AI Magazine - August 2021

TheDIGITAL FUTURE ofSpace Exploration

August 2021 | aimagazine.com

Revolutionary R&D: AI, Mixed Reality, and Digital Twins

Ron Thompson, NASA’s Chief Data Officer and Digital Transformation Officer shares the space agency’s vision

Machine Learning: NLP: Fracking for Data AI Applications: The Future of Cloud and AI

Artificial Intelligence companies to watch

FEATURING:

Saphyre

CPQI

Never miss an issue!

+ Discover the latest news and insights about Global AI...

JOIN THE COMMUNITY

The AI Team EDITOR-IN-CHIEF

PADDY SMITH DEPUTY EDITOR

LAURA BERRILL

PRODUCTION DIRECTORS

GEORGIA ALLEN DANIELA KIANICKOVÁ PRODUCTION MANAGERS

EDITORIAL DIRECTOR

SCOTT BIRCH

OWEN MARTIN PHILLINE VICENTE JENNIFER SMITH

CREATIVE TEAM

PRODUCTION EDITOR

OSCAR HATHAWAY SOPHIE-ANN PINNELL HECTOR PENROSE SAM HUBBARD MIMI GUNN JUSTIN SMITH REBEKAHBIRLESON DUKE WEATHERILL JORDAN WOOD

JANET BRICE

VIDEO PRODUCTION MANAGER

MEDIA SALES DIRECTOR

DIGITAL VIDEO PRODUCERS

SALES AND MARKETING DIRECTOR

KIERAN WAITE

SAM KEMP EVELYN HUANG TYLER LIVINGSTONE HABBIE AMOS JACK NICHOLLS MARTA EUGENIO MARKETING MANAGER

ANDREW STUBBINGS PROJECT DIRECTORS

KRIS PALMER MIKE SADR BEN MALTBY TOM VENTURO CRAIG KILLINGBACK

JAMES WHITE

JASON WESTGATE MANAGING DIRECTOR

LEWIS VAUGHAN

CHIEF OPERATIONS OFFICER

STACY NORMAN PRESIDENT & CEO

GLEN WHITE

FOREWORD

The washing machine paradox AI, like washing machines, will throw up some surprises you weren’t expecting. We should all manage our expectations.

“The assumption that automation would mean less work was flawed”

AI MAGAZINE IS PUBLISHED BY

Let’s talk about washing machines. Because… well, we’ll get to that later. But for now, we can discuss a trend no one expected from modern household appliances.

When they were invented, they were sold as laboursaving devices. And it was a believable line. Putting a load of clothes in the twin tub, shaking in a bit of Daz and hitting ‘start’ is easier than operating a mangle (I guess, I’ve never used a mangle).

But here’s what really happened: people (women, really) who had a washing machine, dishwasher, vacuum cleaner found they were spending longer on household tasks than previously.

The data are arguable, but the concept is not. The assumption that automation would mean less work was flawed. There was still an operator who had to load and unload, so the labour was only reduced, not gone. Expectations for how often to clean clothes and carpets also shifted: an increase. And there were further additions to the workload – the red sock in the white wash that introduced hours of stain removal; the blocked filter, the malfunction requiring specialist servicing. I don’t think these problems were prevalent with a good old mangle (again, I’m assuming).

Thing is, AI is just the same. It could and should make things better and easier. We still use household appliances for a reason. But it’s not sci-fi: manage your expectations or prepare to be surprised.

Do get in touch with your mangle experiences please.

PADDY SMITH

paddy.smith@bizclikmedia.com

© 2021 | ALL RIGHTS RESERVED

aimagazine.com

3

THE ULTIMATE TECHNOLOGY, AI & CYBER EVENT SEPTEMBER

14th - 16th STREAMED LIVE FROM TOBACCO DOCK LONDON A BizClik Media Group Brand

Confirmed Speakers Include: Danny Attias

Chief Digital & Information Officer Anthony Nolan

Marcell Vollmer

Chief Digital Officer Boston Consultant Group

Kate Maxwell

Chief Technology Officer Microsoft

Renata Spinks

CISO United States Marine Corps

Scott Petty

Chief Technology Officer Vodafone

EARLY BIRD TICKETS HERE

Creating Digital Communities

CONTENTS

Our Regular Upfront Section: 8

Big Picture

10 The Brief 12 Global News 14 People Moves 16 Timeline:

The history of IoT 18 Trailblazer:

Demis Hassabis 20 Five Mins With:

Jonathan Epstein

24

NASA

The digital future of space exploration

38

AI Strategy

Revolutionary R&D: AI, Mixed Reality, and Digital Twins

46

Saphyre

Democratising trading through agnostic technology

82

AI Applications

The Future of Cloud and AI

58

Machine Learning NLP: Fracking for Data

88

Data & Analytics Bank on the Right Outcome

Following the Data Science

Leading Provider of Digital Transformation and Managed Services

66 CPQI

Bank on the right outcome and future proof your institution

96

Top 10

AI Companies

BIG PICTURE

Forgetting Facebook Menlo Park, US

Facebook has developed machine learning models that can ‘forget’ information that is no longer relevant, as the human brain does. It’s a big step for AI, which currently collects and stores data indiscriminately.

Selective data would have cost efficiencies for businesses as well as possibly improving historical data mining capabilities.

8

August 2021

aimagazine.com

9

THE BRIEF “Across industries, from utilities and the public sector, to proptech and financial services, there is growing interest and investment in edge computing” Jack Lawton

Data science principal at data analytics specialist at UK-based Aiimi 

BY THE NUMBERS Future employment

50%

Number of people who think roles and skills needed in the next 10-15 years in their industry are impossible to predict

85%

Estimated number of ‘job concepts’ in 2030 that do not yet exist

READ MORE

“Mature organisations are starting to change their strategies and tactics in how they organise their data people and how they build their data science platforms” Andrew Morgan

Director of Data at 6point6 and author of ‘Mastering Spark for Data Science  READ MORE

“In every sector you can find someone peddling an AI system to make decisions, across businesses, government, charities, education, health and the military”

Did you know?

Spend on marketing automation tools is expected to increase to $25.1 billion by 2023, up from $11.4 billion in 2017 READ MORE

Salesforce’s Einstein Relationships Insights AI platform delivers more than 80 billion AI-powered predictions per day READ MORE

Dr Garfield Benjamin

A Postdoctoral Researcher at Solent University  READ MORE

10

August 2021

The chatbot market is expected to reach $1.2 billion globally in less than 10 years READ MORE

 TRAFFIC HAZARD REPORTING AI is being used to predict when road traffic incidents are most likely to occur. The software, which requires information from Android and Apple phones has collected as many as 4,500 observations in a month on a trial on the A14 in Cambridgeshire, UK.

What is data fluency?

010001111000011100101000?

I’m sorry, I don’t speak binary.

Well, you’re missing out. ‘Data fluency’ is the new catchphrase of the wise corporate mind.

Don’t we have computers for that?

Yes, and that’s the point. With the explosion of data as a corporate asset in recent years and the impending boom as AI models get better at data mining, we’re about to be positively swimming in the stuff.

Is it lovely once you’re in?

Probably, if you’re data fluent. The World Economic Forum estimates automation will displace 75 million jobs by the end of 2022 and create 133 million others. Most of the new ones will require data skills.

And that’s data fluency?

Exactly. Companies need to upskill workers in data immediately to stay ahead of the stampede for data talent that lies ahead.

Those that train become data fluent.

Now you’re talking my language.

 IMAGINATION AI While humans can easily put together mental images of things they have never seen, AI has struggled with the recombinative process of applying, say, shape and colour to an artefact. A USC team has developed AI that allows computers to ‘imagine’.  DEEP FAKE DOCUMENTARIES A documentary maker who used a deep fake company to replicate the voice of Anthony Bourdain has been censured by critics, including the dead celebrity chef’s ex-wife.

Filmmaker Morgan Neville described the technology as a “modern storytelling technique”.  AI DIAGNOSTICS Healthcare has been a primary sector for the AI industry’s affections, with promises of rapid diagnoses, drug delivery and public health analysis. But a new report published in Science argues that the benefits of AI in health are oversold.

W A Y U P

AUG21

W A Y D O W N

aimagazine.com

11

GLOBAL NEWS

1

CALFORNIA, US

AI chip floorplans Previously defying attempts at automation, chip floorplanning – the process of deciding where elements of memory and logic are located on a microprocessor – has been achieved by AI with results that superior to humans’ efforts. In a paper in Nature, researchers from Google Brain reported that the machine learning approach took just hours, versus weeks or months it would take a human.

12

August 2021

2

UNITED KINGDOM

Big Brother is driving with you UK-based MiX Telematics has upgraded its AI video telematics service to include details of driver fatigue, phone use, distraction, smoking and seat belt use. The new platform can live stream driver behaviour, giving fleet managers more visibility. Richard Adams, Sales Director at MiX Telematics Europe, said: “We have demonstrated that the return values of this solution are very high for our customers.”

3

INDIA

Loans worry An explosion in AI-based loan apps in India is creating problems with transparency and accountability. Many young people in the country have turned to lending apps as smartphone usage has boomed in the country, yet some of the new fintech lenders lack effective channels of complaint, leading to credit ratings being damaged. The algorithms are also said to disproportionately exclude marginalised groups, including women.

4

CHINA

Senate bill row China has expressed “strong dissatisfaction” at a US Senate bill – the Innovation and Competition Act – aimed at boosting American technology and manufacturing. The National People’s Congress said: “This bill seeks to exaggerate and spread the so-called China threat to maintain global American hegemony, using human rights and religion as excuses to interfere in China’s domestic politics and deprive China of its legitimate development rights.”

aimagazine.com

13

PEOPLE MOVES ADAM SELIPSKY FROM: TABLEAU TO: AMAZON WEB SERVICES WAS: CEO NOW: CEO Adam Selipsky is the incoming CEO of Amazon Web Services (AWS), the world’s most comprehensive and broadly adopted cloud platform. Having previously led AWS marketing, sales, and support for 11 years, from 2005-16, Selipsky helped launch and grow AWS from a start-up into a multi-billion dollar business. Prior to rejoining AWS in 2021, Selipsky was most recently president and CEO of Tableau Software. He led Tableau through its acquisition by Salesforce, in what was the third largest software industry acquisition at the time.

Selipsky will take the office of Andy Jassy, who is succeeding Jeff Bezos as CEO of parent company Amazon.

“Adam brings strong judgment, customer obsession, team building, demand generation and CEO experience to an already very strong AWS leadership team,” Jassy said.

14

August 2021

ALEXA ANTHONY FROM: MAGIC AI TO: THRUWAVE WAS: CEO NOW: BUSINESS OPERATIONS

DAMON FLETCHER

Alexa Anthony co-founded and ran Magic AI from 2017 to 2020 but the horse monitoring software failed to draw further investment and was shuttered. She has been picked up by ThruWave, a University of Washington spinout that uses millimeter waves to see through packaging. The technology has applications in ecommerce and warehousing, and the startup raised $6.4 million last year. She said: “There is a computer vision and hardware component, an industrial market, incredible talent on the team, great progress with customer contracts in place, and early revenue.”

FROM: TABLEAU TO: DATAROBOT WAS: CFO NOW: CFO Tough times at Tableau as it loses its second senior exec just after its CEO left for AWS’ shores (see above). CFO Damon Fletcher had been at the company for seven and a half years before his departure for DataRobot, an enterprise AI platform. Before joining Tableau, Fletcher worked at PwC for over 11 years.

After beginning his new role Fletcher said: “I spent the week with the senior leadership team at DataRobot and I am extremely excited about how we are helping our customers drive huge ROI.” aimagazine.com

15

TIMELINE THE HISTORY OF IOT The Internet of Things (IoT) is not nearly as new a concept as you might think, dating back – in concept form, at least – to the 1800s

16

1830s

1980

Machine to machine communication The essence of IoT is one machine speaking to another, a concept that arrived with the telegraph in the 1830s.

We can skip ahead a few chapters here: radio (1900), the arrival of computers (1950s) and of course the internet (1962, if you count early DARPA development).

You want a Coke with that?

An early example of modern IoT was a Coca-Cola vending machine at Carnegie Melon University in Pittsburgh, Pennsylvania. Local programmers could connect to the machine to establish if drinks were available and – if so – if they were chilled.

August 2021

1999

2000s

2017

Kevin Ashton Kevin Ashton from MIT described modern IoT at a presentation to Procter & Gamble. He said: “If we had computers that knew everything there was to know about things, using data they gathered without any help from us, we would be able to track and count everything and greatly reduce waste, loss and cost. We would know when things needed replacing, repairing or recalling and whether they were fresh or past their best.”

Toy time As industry began to experiment with more IoT, the trend bled into consumers’ lives with connected toys, cameras, doorbells and appliances.

The term IoT morphed from meaning ‘anything connected to the internet’ to ‘anything ancillary connected to the internet’. RFID and QR codes began to close the space between the connected world and ‘meatspace’.

IPV6 The upgrade of the internet to its current protocol, IPV6, anticipated a huge demand for internet addresses.

According to Steve Leibson at the Computer History Museum in Mountain View, California, “We could assign an IPV6 address to every atom on the surface of the earth, and still have enough addresses left to do another 100-plus earths.” IoT, your future is safe.

aimagazine.com

17

TRAILBLAZER

Demis Hassabis Job Title: Founder Company: DeepMind Superhero of AI

D

emis Hassabis, dubbed ‘the superhero of artificial intelligence’, is a man with a deep understanding of the mind.

A classic overachiever, he started playing chess aged four, achieving master standard at 13. With his chess winnings he bought and modified his first computer, a ZX Spectrum.

Meanwhile, he continued to study, completing his A-levels at 16 to gain a place at Cambridge University.

Before going up, Hassabis took a gap year, eschewing backpacking in favour of working for a computer games company, Bullfrog. At 17, he codesigned and led the programming of Theme Park, an award-winning and influential simulation game that sold more than 10 million copies. Having completed his gap year, he bought a Porsche and turned up at Fresher’s Week in it.

Hassabis returned to the gaming industry after his studies and was instrumental in pushing the boundaries of in-game AI. But he would return to academia in 2009 when he went to University College

18

August 2021

Google paid

400m

for part of DeepMind in

2014

London to study for a PhD in cognitive neuroscience. His papers – notably one establishing a link between constructive and reconstructive processes in the brain – were considered major breakthroughs in the field.

Inevitably, Hassabis would return to the field of AI armed with a deeper understanding of human intelligence and how its application might be used in technology. He founded DeepMind in 2010 with the intention of “solving intelligence” then using intelligence “to solve everything else.”

In an interview with The Verge in 2017, Hassabis explained that while the link between AI and neuroscience was important, the subject was too complicated to transfer directly to use-case scenarios common to AI solution finding.

“If you’re an AI expert today and you have no neuroscience background at all and you try getting into it, it’s quite daunting,” he said. “I think there’s something like 50,000 papers a year — I can’t remember the exact number — published in neuroscience. So there’s a huge body of work to try and make sense of, most of which is not going to be relevant to AI, meaning you’re looking for nuggets of crucial information in a huge haystack.”

Hassabis company DeepMind was partially acquired by Google in 2014 for $400 million. It has been incorporated into Google Health and was behind the creation of AlphaGo, the program that defeated Go world champion Lee Sedol. It beat him 4-1 at the famously complex game.

“There’s a huge body of [neuroscience] work, most of which is not going to be relevant to AI. You’re looking for nuggets of crucial information in a huge haystack” Demis Hassabis aimagazine.com

19

FIVE MINS WITH...

JONATHAN EPSTEIN Brewco is bringing AI to SEO to bring new dynamics to a previously static field. Its CEO tells us what’s changing in this competitive market

Q. TELL US ABOUT BREWCO

» Brewco was formed by a team

of industry veterans that have previously introduced other AI-based marketing platforms. They are the official reseller of Market Brew, an AI platform for precision search engine optimization (SEO), transforming the way companies and their agencies can optimize their pages to rank higher in search results. Organic search results drive 53% of all site visits, so having a competitive advantage in SEO makes a huge difference in a company’s revenues and profitability.

Q. WHAT IS MARKET BREW?

» Market Brew builds highly

“ AI IS ENTERING THE SEO MARKET” 20

August 2021

accurate models of how search engines weigh a wide range of factors to determine rankings for a keyword. Each search has different weights, so no rules-based approach provides these insights.

From an AI standpoint, Market Brew uses particle swarm optimization to evolve the weights of its own search engine until it delivers a search engine results page (SERP) that is highly correlated with Google’s (or Bing etc) Once this model is developed, it scores your website and others that compete for rankings, identifies the statistical gaps you should address, and provides precise recommendations on closing these gaps.

“ SEO FOR VOICE, VIDEO, AND CLOSED ENVIRONMENTS LIKE AMAZON AND SOCIAL MEDIA IS ALSO A HORIZON THAT LEADING-EDGE SEOS ARE BEGINNING TO PLUMB” Q. HOW IS SEO CHANGING?

» AI is entering the SEO market.

There are a variety of solutions that handle different aspects of the SEO process, from addressing how to show up in Google’s Q&A search results, build better schema, and develop backlinks.

Google is releasing its first new algorithm in some time – Core Web Vitals. As a result, searches will now also consider page performance measures such as load times and when users can first interact.

Fortunately, Market Brew’s search engine modeling approach will be able to track precisely how much these factors matter.

Q. WHAT DOES THE FUTURE OF SEO LOOK LIKE?

» Google’s SERPs continue to

evolve, and there are now many different ways a site can rank, in addition to the core site rankings.

As a result, we assume they will continue to refine the space on the SERP in ways that will keep SEOs on their toes.

Doing SEO for voice, video, and closed environments like Amazon and social media is also a horizon that leading-edge SEOs are beginning to plumb.

Overall, though, the black box AI nature of search engines will make the job of SEOs more difficult unless they have AI on their side as well. aimagazine.com

21

Top 100 Leaders in Technology September 2021 To be announced at the Technology & AI LIVE Event NOMINATE NOW

A BizClik Media Group Brand

Creating Digital Communities

NASA

SpaceX Crew-2 Launch

The digital future of space exploration WRITTEN BY: PADDY SMITH PRODUCED BY: MIKE SADR

aimagazine.com

25

NASA

The US space agency is part way through a digital transformation that it hopes will thrust it towards the future. Its chief data officer talks about why that matters

N

ASA is a young federal agency with fewer than 100 birthdays under its belt. That modernity ought to be a launchpad fit for digital transformation.

But the organisation that first put a man on the moon and has continued to lead space exploration in the intervening 60-odd years is travelling light when it comes to legacy tech baggage.

Ron Thompson is NASA’s chief data officer and deputy digital transformation officer. He explains that, while NASA is a relatively ‘new’ agency, many of its operational and organisational models were passed down from existing agencies at its inception. “The way the agency works is really routing physical paper artifacts, for different activities and a lot of that culture was based around routing mail,” Thompson explains. “It’s mail codes how the organisations operate and how systems are designed to support the agency” It’s little surprise then that Thompson is an advocate for changing cultural and workforce norms in order to usher in a digital transformation programme. “It’s really looking at the enterprise and working holistically, finding areas of opportunity to share and breaking down those mail code barriers and the system barriers that were in place based on those mail codes, interoperability is the name of the game.”

The digital transformation team Thompson is part of is “small but mighty”, more akin to a university than a big federal agency such as the Department of Defense.

NASA Astronaut Edward White during first EVA performed during Gemini4 flight

aimagazine.com

27

NASA

Astronauts Megan McArthur and Randy Bresnik at KSC

NASA itself is structured into three domains: human exploration, science and research, and aeronautics. “If you look at how those cultures have derived, what we’re seeing is world class instantiations of things, but they were never designed to operate and interconnect and interoperate as efficiently and effectively as possible, hence the opportunity to transform.”

That means Thompson and his small but mighty team of five need to look not at how NASA’s systems are operating but at how NASA itself is operating. “That’s really the core tenet of our work, to examine the operating model and look for areas of transformation. We’re not doing digital transformation for digital’s sake. We’re doing it to ensure we have mission outcomes that absolutely advance the work.”

There is a realisation from senior leadership that the agency cannot continue to operate under its current constraints and for that Thompson’s team has been charged with 28

August 2021

streamlining the complexity of missions.

Those missions have become increasingly diverse, with trips to Mars and resupply cargo to the International Space Station continuing through Covid. Meanwhile, the agency’s resources are coming down.

For answers, Thompson has looked to the private sector to see what can be

“ We’re not doing digital transformation for digital’s sake. We’re doing it to ensure we have mission outcomes that absolutely advance the work” RON THOMPSON CDO, NASA

NASA

learned. It is, Thompson admits, “very much ahead of the federal sector in the chief data officer role and where that sits.” To that end, Thompson sees opportunities to increase hindsight, insight and foresight. Part of that is changing the hierarchical structure of the agency from top-down to a combination with bottom-up.

“The young workforce demands transparency and they demand to have information available to them to advance their work. The opportunity is right in the middle. That’s why in our culture we really have to explain what we’re trying to solve.

For us, digital transformation is unlocking the door for a massive future. But we also need to gravitate around why we need to change and how.”

All this must happen while missions continue to launch. Nothing can stop during the transformation, which will be integrated into the critical day-to-day work of NASA. Thompson has done his reading. “If we look at the studies, we know that 70 percent of these efforts fail, so we do not want to be in that category. We want to learn and iterate but we also recognise that is high risk. When you prototype things that are different, it’s high risk and great value, great reward, great

Jobs across the United States

$23.2 bn Annual budget for 2021 fiscal year

18,000 Workforce of civil servants

TITLE: CDO INDUSTRY: AVIATION & AEROSPACE LOCATION: WASHINGTON DC

EXECUTIVE BIO

312,000 +

RON THOMPSON

Ron has more than 30 years of experience in leadership roles spanning the federal government and public sectors. Throughout his career, he has been responsible for organization transformation, data design, innovation, engineering, operational management and governance of complex IT solutions.

Ron’s breadth of hands-on experience in organizational design and transformation will bring valuable perspectives in partnerships and creative problem solving.

Currently Ron is serving as the Agency Chief Data Officer (CDO) and Deputy Digital Transformation Officer (DDTO), reporting to the Deputy Associate Administrator at NASA Headquarters.

Ron previously served Associate Chief Information Officer (ACIO) Transformation and Data Division (TDD) & for the NASA Office of the Chief Information Officer (OCIO).

NASA

Connect your digital business securely.

From user to application in a cloud-first world.

Learn more:

cisco.com/go/networking

Cisco: reaching for the stars with NASA

and the foundational importance of data to a digital economy. “We really pride ourselves on the idea that it’s built-in, not bolted on,” Scheaffer says.

Cisco’s Jeff Scheaffer explains how the company is helping the US space agency digitally transform for the future of space exploration.

Cisco supports NASA’s six core ‘thrusts’: data and insight centric; focused on collaboration; culture and workforce; “modelbased anything”; AI and machine learning; and process transformation. These thrusts are “completely integrated” into NASA’s transformation architecture. Scheaffer describes Cisco’s programme for NASA as “laser focused” on achieving NASA’s mission.

In common with NASA, Cisco invests heavily in R&D, spending $6.3 billion annually on new networking and security technologies.

It’s a shared cultural asset that will help the company in its endeavors to assist the US space agency’s digital transformation – and future missions.

Jeff Scheaffer is VP Product Management & Strategy at Cisco, focusing on softwaredefined networking (SDN) transformation.

These networks, Scheaffer says, “are much more flexible environments – they’re better suited to ensuring application experience because software defined networks are more rapidly and flexibly reconfigurable, which improves our agility and our ability to respond to application quality of service and security issues.”

Data is – obviously – core to the joint missions of NASA and Cisco. As a networking provider, Cisco has a deep understanding of its AI stack

“It’s been a phenomenal partnership as we work with NASA to identify the next set of needs to drive the next generation of product innovation cycles. We’re thrilled to be partners and to support NASA and all of the collaboration and the benefit that comes from that partnership.

“We’re really excited in partnership with NASA to be able to focus on using that investment to help us to create comprehensive, flexible, digitised infrastructure – real-world and real time analytics – in order to provide better experiences, agility, and trust.”

LEARN MORE

NASA

GRACE Follow-On Satellite

DID YOU KNOW...

NASA’S CLIMATE MISSION

32

Sitting in a tin can far above the world is, it turns out, a pretty good place to survey changes to Earth’s climate, a pivotal mission for our times. Not only does the climate need monitoring, but policymakers need to know which of their schemes are working. For that NASA is working with Boston Consulting Group to work out how to collate the data it’s collecting. “We want to start a pilot, probably in the next six months, to really take a look at the climate data NASA has, while bringing in other data sets from other organisations [such as the European Space Agency, ESA],” says Ron Thompson, chief data officer (CDO) and deputy digital transformation Officer at NASA. “That way we can absolutely show policymakers how the climate policies make an impact. We can take the information we collect today and do ‘what if’ trade-offs on climate. We’re working with Gavin Schmidt, our acting climate change lead at NASA, on coming up with a prototype to address some of these questions. That will be in the cloud, using publicly available datasets.”

August 2021

upside, but it is something different. The opportunity is to really integrate this across the agency.”

One area NASA is looking to improve is in process transformation – financial, travel and so on. The processes must be repeatable across all divisions of the organisation and easily understood. Thompson is using robotic process automation (RPA), which he describes as “the most mature area where we’re seeing value in digital transformation.”

The value comes in freeing up time to work on “more meaningful, challenging work”.

But technology at NASA doesn’t just mean back office functions. “We’re looking at doing a digital twin for our Orion mission, the next man mission to the moon. We want to set that up to test our common models and use agency-wide standards using modelbased systems and engineering across the agency.” That digital twin can be used to analyse safety for mission insurance. AI is also

Efficient Descent Advisor, Simulaiton Number 2, ATC Lab N-257

NASA

being employed to ferry information securely between systems in a cloud network and to improve astronaut communication using natural language processing (NLP).

“We’re setting up our data-driven decision lens on how we visualise our data across NASA. That’s true for mission data. That’s true for back office data. That’s true for HR data.

We’re setting up that visualisation platform that is a consistent look and feel but it also links into our data sources on the back end consistently through APIs.” There’s also an enterprise data platform with an analytics capability to deliver “powerdriven data decision making” enhancing veracity, quality and consistency. This platform also holds the key to partnership collaboration with digitally native communication being a core goal, including virtual and augmented reality solutions for engineering projects.

“That,” says Thompson, “is a real advantage of NASA because we’re very collaborative.

I like to describe it as a guild. We have multiple guilds through the organisation coming together and working on solving a specific problem. We call it ‘swarm innovation’ – it’s what we’re using to bring communities together, to share and advance their work in an agency-wide perspective.”

It sounds impressive, but these high-value projects represent just six months or so of launches. Thompson estimates that there are around 35 more for next year, bridging process transformation, AI/ML and data collaboration. “We’re not doing digital for digital’s sake – we’re actually taking a look at our business model and how NASA operates and impacting mission value.”

The end goal? “What we’re gravitating towards is improving our pace of delivery,” says Thompson. “If you look at pace of delivery for launches, pace of delivery for robotic missions, pace of delivery for science and research studying climate,

“ Digital transformation is unlocking the door for a massive future” RON THOMPSON CDO, NASA

aimagazine.com

33

NASA

Orion Forward Bay Cover at Lockheed Martin near Denver

pace of delivery for everything we do in the organisation.

“The big play here is for digital transformation to be the integrator across the agency, to take that enterprise perspective, to see where the common functionalities and common pain points exist and to look at where we can bring enterprise solutions to resolve these. People are getting behind this and we have a goal of increasing our pace of delivery over the next three to five years by about 25 per cent.”

End goals in digital transformation tend to be speculative. There’s a notional idea 34

August 2021

of progress with a view to revising systems and processes as new technologies and ideas emerge. But NASA is an organisation that prides itself on having big dreams with definitive end points and its digital transformation is no exception. “We do not want this to be an effort that extends into perpetuity,” says Thompson. “We’re successful when everything the IT shop does is baked into everything. We need to remove the barriers of entry for our work. We’re putting ourselves out of business. We need to change that mindset and culture so that everything we do is inculcated and infused

NASA

NASA: the digital future of space exploration

“We’re looking at doing a digital twin for our Orion mission, the next man mission to the moon. We want to set that up to test our common models and use agency-wide standards using model-based systems and engineering across the agency” RON THOMPSON CDO, NASA

aimagazine.com

35

NASA

Orion ITL Orion Simulator

DID YOU KNOW...

ORION – NASA’S NEXT MANNED SPACE EXPLORATION PROGRAMME

36

For the first time in a generation, NASA is building a human spacecraft for deep-space missions that will usher in a new era of space exploration. A series of increasingly challenging missions awaits, and this new spacecraft will take us further than we’ve gone before, including to the vicinity of the Moon and Mars. Named after one of the largest constellations in the night sky and drawing from more than 50 years of spaceflight research and development, the Orion spacecraft is designed to meet the evolving needs of our nation’s deep space exploration program for decades to come. Orion deep space exploration missions, coupled with record levels of private investment in space, will help put NASA and America in a position to unlock the mysteries of space and to ensure this

August 2021

nation’s world preeminence in exploring the cosmos.

Orion will serve as the exploration vehicle that will carry the crew to space, provide emergency abort capability, sustain astronauts during their missions and provide safe re-entry from deep space return velocities. Orion missions will launch from NASA’s modernised spaceport at Kennedy Space Center in Florida on the agency’s new, powerful heavy-lift rocket, the Space Launch System. On the first integrated mission, Artemis I, an uncrewed Orion will venture thousands of miles beyond the Moon over the course of about three weeks. The mission will pave the way for flights with astronauts beginning in the early 2020s.

Source: NASA

Orion Exploration Flight Test (EFT)- 1

right out of the gate. We need to be the instigators. We must align.”

Whether internally or working with partners, Thompson sees this “alignment opportunity” as a way to stop multiple parties working independently on what is, essentially, the same problem, instead coming together and achieving more.

“Maybe it’s smart procurement, maybe it’s smart designs, maybe it’s the future of work. And what does NASA look like in a hybrid environment where we have a virtual presence and a physical presence onsite?

What does that experience look like from the user perspective? Is it cumbersome?

We’re thinking about these things and how they interoperate and interconnect. We’re thinking there is a time – we’re not sure when yet – where this group is, everything this group does, is just naturally, organically flowing with everything we do.”

That’s not to say NASA doesn’t have some timelines in mind. Thompson thinks there are milestones that are “absolutely doable” in three to five years. “I see a world where we have online systems that can tell us how all of our projects are working, how everything is integrated and interconnected, how we have real time access to a decision lens for project reviews, for safety reviews, for performance reviews. Everything we do.

It’s very integrated.”

“ We’re successful when everything the IT shop does is baked into everything”

NASA

It’s an ambitious deadline given a “paper mentality” still exists in pockets of NASA. But Thompson thinks breaking down barriers, moving away from the mail code system and introducing a culture of collaboration and “a single source of truth” is entirely possible within the timeframe.

“I’m seeing a modern, immersive world,” Thompson says, “that can use our technology. It’s crossing the barriers. We need to open with transparency that has the right controls and security around them, but that information has been liberated across the organisation and is accessible and absolutely helps with our hindsight, insight and foresight. At that point, we’re ready to answer all the questions.”

RON THOMPSON CDO, NASA

aimagazine.com

37

AI STRATEGY

Revolutionary R&D: AI, Mixed Reality, and Digital Twins

38

August 2021

AI STRATEGY

Virtual, augmented, and mixed reality enable companies to develop, test, and refine innovative prototypes

D

WRITTEN BY: ELISE LEISE

ata is everywhere. But humans aren’t good at analysing big data, and all the data in the world isn’t much use if you can’t make sense of it. Artificial intelligence, on the other hand, isn’t deterred by the vast, complex oceans of intelligence we’ve created. In fact, ‘companies that invest in industry-specific machine learning research will generate extremely valuable intellectual property’, said Foteini Agrafioti in the Harvard Business Review—suggesting that going forward, humans will adopt AI to help analyse, sort, and solve their wicked problems in R&D. As Isaac Newton might have said had he witnessed this shift, AI allows us to stand on the shoulders of giants. We possess billions of bits of data from the achievements of the 21st century—self-driving vehicle performance, Amazon customer feedback, insurance payback rates. Yet it’s only with artificial intelligence and machine learning that we’ve identified which patterns are actually important.

How AI is Changing the Face of R&D Until now, researchers spent stupendous amounts of time trudging through data swamps. With the sheer amount of scientific literature available, they spent months searching for the papers that held the exact information they sought. But no longer.

Artificial intelligence reduces the time R&D aimagazine.com

39

Enabling educators.

Empowering students.

Explore how we accelerate student discovery, learning and innovation with our Digital Education 3D Experience.

EXPLORE THE 3D EXPERIENCE

AI STRATEGY

“ Companies that invest in industry-specific machine learning research will generate extremely valuable intellectual property” FOTEINI AGRAFIOTI HEAD, BOREALIS AI

Pharmaceuticals teams need to develop their prototypes.

For example, natural language processing (NLP) gives computers the ability to analyse, categorise, and select relevant research, allowing humans to focus on what they do best: brainstorm, design, and collaborate.

How does NLP do this? Starting with unstructured text, which could include anything from research PDFs to LinkedIn posts, the programme can search for and suggest relevant information. Humans may not be able to sort through the 2 million scientific articles published annually,

By 2024, AI in pharma R&D is projected to reach US$204bn, primarily in drug discovery and genomics. As with the COVID vaccine, every day counts when you’re waiting for a jab. AI systems can help build doctors examine root causes, researchers pick out vaccine mutations, and governments track and roll out vaccine programmes. ‘It’s a way that we can avoid these lengthy Phase III trials’, said Matthew Putnam, CEO of Nanotronics. ‘Everything that’s going on right now is so incredibly urgent’.

aimagazine.com

41

AI STRATEGY

Manufacturing It can take two to five years to design an electric vehicle or an integrated circuit.

Artificial intelligence can help engineers select the optimal build materials, react to global supply shortages, and fix errors when they’re least expensive—in the simulation phase. According to the University of New Mexico, design decisions determine roughly 70% of a product’s manufacturing costs.

This means that R&D can lower costs while speeding up the time it takes to go to market.

but NLP can. In pharmaceutical R&D, AI scans scientific literature and patient data to identify potential trial compounds, proposes which molecules to target, and suggests test patients.

Furthermore, much of what we consider innovation depends on making predictions and gathering insights from past data.

• AI. Helps R&D spot patterns that humans might miss, suggesting unexpected ideas.

• Machine Learning. Classifies data, predicts trends, and optimises prototype designs.

IBM, for example, used AI to evaluate 20,000 potential electrolyte compounds for a new electric battery. Without AI, the job would have taken five years. Instead, IBM finished it in nine days.

42

August 2021

Virtual, Augmented and Mixed Reality In addition to data analysis, AI aids with design. By building detailed visualisations at high fidelity, R&D teams can delay using physical materials until the final stage of the development process. This is critical in the automotive industry, where test cars can cut into your company’s budget. Back in 2017, Ford recognised the potential of mixed reality and used it to prototype future vehicles. Fast-forward to today, and Volvo has gone even further, using its Varjo XR-1 Developer headset to model its vehicles in real-time—even recreating the physical sensation of what it’s like to drive.

These 3D environments free us from some of the constraints of our physical environment. R&D teams can test their

“ It’s an approach and way of working, enabled by connecting the physical and digital worlds, that will fundamentally change our society, business, economy, and environment” SIMON EVANS

DELIVERY TEAM LEAD FOR BRITAIN’S NATIONAL DIGITAL TWIN PROGRAMME

designs in costly or unlikely real-world scenarios, such as simulating auto crashes or natural disasters without any fallout. Tesla engineers can model electric batteries even when critical minerals are in short supply.

And that’s just the tip of the iceberg.

In recent years, increasingly advanced computing has allowed us to take on more complex simulations. Fugaku’s supercomputer enabled manufacturers to use complex CFD (computational fluid dynamics) simulations—fast, accurate models that reduce design time. These simulations allowed original designs to be fed into simulation software, allowing engineers to pre-test critical safety and performance features prior to build. Some even took simulation technology one step further... aimagazine.com

43

AI STRATEGY

Digital Twins If a simulation helps us virtually model the real world, a digital twin helps us receive feedback from it. A ‘representation of an asset’, it’s essentially a digital copy of a physical prototype. For example, you might build a robotic arm for your assembly line. You put sensors on it. Now, you can send real-time data relating to accuracy, efficiency, and downtime back to its digital copy. With this type of data, your digital simulation can much better approximate the real deal.

Though digital twins have many applications, you see this approach quite often in the automotive industry. Siemens, for example, developed its PAVE360Simcenter Prescan simulation tool and digital twin technology to test how cars would operate in the real world. After accepting CAD files, its R&D simulation recreated a granular representation of each vehicle part: chassis, safety sensors, semiconductors. Without putting a car on the road, this is the closest we can get to This is the closest you can get to modelling real-world performance. Said Simon Evans, Delivery Team Lead for Britain’s National Digital Twin Programme: ‘It’s an approach and way of working, enabled by connecting the physical and digital worlds, that will fundamentally change our society, business, economy, and environment’.

The Future of AI in R&D Artifical intelligence in R&D is just getting started. Between 2020 and 2027, experts predict that the global AI market will grow by 42.2%. ‘It makes jobs far more satisfying because we automate 80% of the work’, said Craig Civil, Head of Data Innovation, R&D, and Analytics at Lloyd’s.

44

August 2021

“ It makes jobs far more satisfying because we automate 80% of the work” CRAIG CIVIL

HEAD OF DATA INNOVATION, R&D, AND ANALYTICS, LLOYD’S

AI STRATEGY

Insurance R&D teams use AI to identify patterns in user data, customise pricing models, and understand customer behavior. As BusinessWire reports, satellite and aerial imagery can help analyse property data.

An example: Arturo, a company that helps insurance companies with predictive analytics, uses deep learning models to provide unique property data in as little as five seconds. In another use case, NLP analyses financial statements and investor presentations. Both help insurance companies understand what their clients will pay for—and what they’ll pay back.

‘The remaining 20% is made up of one-off questions that only experienced teams can answer’. What he’s hinting at is an integration of the broad insights offered

by AI and the specialised knowledge of R&D experts.

While the benefits of a R&D future built on AI include a faster time to market, more customised components, and optimised labour, there are those who fear its blind spots—bias, overconfidence, occasional error. Indeed, several organisations have turned to regression testing to ensure that AI software is constrained in scope and unable to make decisions on its own.

Is this right? It may provide oversight, but it remains reminiscent of Vonnegut’s Harrison Bergeron: handicapped by the government to impede his intellectual capabilities.

Fundamentally, the question is this: AI can do the analytical work of R&D—but who should make the final call? aimagazine.com

45

SAPHYRE

46

August 2021

SAPHYRE

DEMOCRATISING TRADING THROUGH AGNOSTIC TECHNOLOGY WRITTEN BY: WILLIAM GIRLING

PRODUCED BY: MICHAEL BANYARD aimagazine.com

47

48

August 2021

SAPHYRE

Stephen and Gabino Roche return to describe Saphyre’s progress as it democratises the pre- and post-trade space through sophisticated yet simple tech

W

hen Stephen and Gabino Roche (President and CEO respectively) at Saphyre spoke with us in 2020, they made it clear that the company was on a mission to overcome risk-aversion in finance and introduce critical innovation to a historically inefficient system. Having realised that silos and poor user experiences were benefitting no one in the pre-trade space, the brothers decided that nothing less than a wholesale disruption of the status quo would satisfy their desire to democratise trading. Saphyre subsequently achieved this ambition with the development of its integrated onboarding platform, but the story didn’t end there.

In a fresh interview, the Roche brothers updated us on what has transpired in the interim, the importance of interoperability, and why customer acquisition will be their primary focus moving forward.

One thing that’s certainly changed has been the COVID-19 pandemic: from not knowing how the situation would resolve in late 2020 to the development of vaccines in Q1 of 2021, Saphyre has weathered the storm and emerged stronger as a result. However, Gabino says, the journey wasn’t always an easy one. “We were a little nervous in the beginning because meeting face-to-face with clients is an important thing when building a business.

Saphyre is a technology company, so it wasn't hard for us to handle going remote, but we were afraid that would create a dip in our client

engagement.” They needn’t have worried; it didn’t. In fact, by utilising video conferencing software and saving travel time between engagements, Saphyre actually managed to create more opportunities than ever. Now, as a degree of normality is gradually restored to life and work in the US, Gabino states that lessons learned regarding remote working’s efficacy for the business will not be forgotten.

The autumn of 2020 brought Saphyre one of its greatest accomplishments to date: partnership with both BlackRock and JP Morgan, who opted to utilise the company’s artificial intelligence (AI) technology to automate account opening for securities services. Stephen makes it clear that important collaborations have continued to be formed, most recently with investment bank BNY Mellon. “Its asset servicing arm is now deploying the Saphyre solution on their platform and working with their counterparties. We’re also working on other partnerships and integrations that you'll be hearing about in the coming weeks and

“ Saphyre is a technology company, so it wasn't hard for us to handle going remote” GABINO ROCHE CEO, SAPHYRE

aimagazine.com

49

SAPHYRE

Saphyre | Bringing Fintech Software Products to Market Fast

months.” Besides this, the company has also been striving to improve the intelligence of its platform and building awareness. Gabino adds, “I think the firms that we're dealing with today understand the value proposition and benefits we bring to pre- and post-trading.”

Since it was founded in 2017, Saphyre has been committed to eliminating manual processes in pre-trade activities. It’s a point on which the company has been “evangelical” because, Stephen says, the

“ We've been trying to get people to think about how to do things in the pre-trade space differently” STEPHEN ROCHE

PRESIDENT, SAPHYRE

50

August 2021

problem is rooted in culture as much as technology. “We've been trying to get people to think about how to do things in the pretrade space differently. Traders want to fix failures or mismatches in post-trade and that’s generally where all the investment goes. Saphyre is telling everybody, ‘Hey, if you just set things up properly in the pretrade space, we can reuse that information during trading and post-trade.’” Essentially, the company believes in a ‘measure twice, cut once’ philosophy that emphasises integrating important protocols and safety measures in the first instance.

Stephen hints that Saphyre’s exploration of the post-trade space is about to yield some truly groundbreaking results for institutions. Until these are revealed, however, Saphyre will continue to break down the silos (or “fiefdoms” in Gabino’s words) that make trading the reserve of a privileged few. “The key difference is that previously we've been talking and evangelising, but

SAPHYRE

STEPHEN ROCHE TITLE: PRESIDENT INDUSTRY: FINANCIAL SERVICES LOCATION: NEW JERSEY, UNITED STATES Stephen has over 20 years of business development and consulting experience in the IT solution space. Working with Fortune 1000 companies and startups.

EXECUTIVE BIO

He spent eight years in the B2C retail environment successfully helping brick and mortar stores gain market share in the New York City region. In 2003 he transitioned his career with AT&T, working exclusively on B2B IT solutions, streamlining complicated global IT infrastructures with 50+ locations from midmarket up to Fortune 1000 clientele.

Always problem solving, Stephen successfully closed large deals concerning emerging technologies, such as 5G and Internet of Things (IoT) with Volvo Cars. In another example, he rolled out AT&T’s state-of-the-art media room, and the renovation of the IT infrastructure for the newly constructed Yankee Stadium in New York City. With Red Bull Racing, he Implemented the latest in racing communication and video technology, along with helping Axis Capital and Wyndham Hotels achieve their strategic IT growth initiatives.

During his AT&T career, he helped establish third-party cloud SaaS offerings for clients such as hosted ERP and ecommerce platforms to realise roughly

US$1mn+ in quarterly earnings. Advising the proper balance between scaling quickly while maintaining operation costs to a minimum has been one of the hallmarks of Stephen’s engagements.

While co-founding Saphyre and managing strategic relationships with the largest financial institutions in the world, Stephen has leveraged his B2B dealmaking experience to build a premiere, professional marketing arm of the upstart fintech firm.

SAPHYRE

2017 Company founded

30

Number of employees

52

August 2021

SAPHYRE

now we're showing a way that this data can be democratised. It's no longer a talking point; we're actually giving people the keys to try out this solution,” says Gabino.

Saphyre’s innovative buy-side portal is being constantly developed to align with and progress this overarching vision. Information that was previously siloed for the community – and therefore difficult to link – has been released through the company’s expanded offerings for the security services space. Now fund accountants, transfer agents, security services, and outsourced middle office teams can now work on behalf of the buy-side firms: “All this interoperability allows those players to collaborate on our platform using the same data in a very intuitive fashion,” explains Gabino. “They can then take that same data and enrich it for the broker dealers and the order management systems.” From a tech (and particularly an AI) perspective, Saphyre never stops evolving. Keen to correct what it views as a decades-long deterioration of the trading tech standard, the company is working to ensure AI fulfills its full potential: mapping data to create an intelligent forecast for how clients will need to use it. “When people say the word ‘to’, do they mean the number two, ‘to’, or ‘too’? Even when you read the word ‘fund’ in the industry, it can be difficult to distinguish what’s being referred to. Those are the nuances that we're actually mapping.”

Achieving this level of automated sophistication meant growing Saphyre’s number of patents from 48 to 67 in a matter of months. “Those patents encompass everything from the logic of how to memorise and understand conceptual relationships between words to an intuitive user interface (UI) where users actually can be lazy without even knowing it,” says Gabino. The result is an AI capable of being highly accurate throughout the entire trading life cycle. aimagazine.com

53

SAPHYRE

SAPHYRE’S SECRET: EMBRACING INTEROPERABILITY When asked if there’s an important industry topic that no one seems to be paying attention to, Stephen opts to highlight ‘interoperability’ as a key trend:

DID YOU KNOW...

“Imagine having an iPhone with all your pictures on it, but you can’t port it to an Android phone. That's quite frustrating; they're your pictures; it’s your data and documents, and these pseudo monopolies perpetuate the problem.

54

“If you truly allow your clients to have a full experience and aren’t afraid of innovation, interoperability can help address that. It allows for the democratisation of data and it's going to provide better margins and superior problem-solving. This is what needs to happen and institutional players need to demand it from their vendors.”

August 2021

Fundamentally, states Stephen, the same system is also technologically agnostic: it is capable of being mapped to a FIX protocol, smart contract, distributed ledger, etc. This provides an important competitive edge in a space where the wholehearted embrace of blockchain still eludes the industry. “One of the big reasons blockchain hasn't been adopted is the fear of adopting a ‘religion’ – a form of standardisation to which everyone has to conform.” Saphyre’s advantage is being able to provide the same operability without such a commitment.

It’s a smart move and one which demonstrates the company’s highly informed and people-centric way of working.

“Partnerships are the most important things for us, besides employees and investors,” Stephen clarifies. “Technology is our second priority, because you have to offer something of value, followed by the patents that come along with it.” Gabino states that many would probably be surprised by the extent that legacy tech mainframes continue to plague finance. Now with a software-tomarket speed that ranks among the fastest in contemporary fintech, Saphyre is primed

“ We're showing a way that this data can be democratised. It's no longer a talking point; we're actually giving people the keys to try out this solution” GABINO ROCHE CEO, SAPHYRE

SAPHYRE

GABINO ROCHE TITLE: CEO INDUSTRY: FINANCIAL SERVICES LOCATION: NEW YORK, UNITED STATES Gabino has over 20 years of experience in building technology solutions for Fortune 500 companies and start-ups from the 1990s Dot-com era till now.

He’s a former McKinsey & Company firm member, where he learned and was focused on delivering products faster to market. This enabled him to take on a role at NYSE as Managing Director of application development to help their startup division ramp towards a US$1bn revenue goal.

EXECUTIVE BIO

Later, he worked as a Senior VP at JP Morgan executing transformation programs in business operations, overseeing technology and operational process initiatives such as delivering a US$40mn product in their custody portfolio, and helping to revamp the Corporate Investment Bank’s (CIB) KYC/ AML operations.

That experience set him up for JPMorgan’s senior management to ask him to take on the Head of Product role at Clarient, a fintech startup consortium put together by JP Morgan, Goldman Sachs, State Street, Credit Suisse, Barclays and DTCC. While there, he oversaw an $80mn budget and transformed the company’s operations, technology, and product with his team

in under six months to meet market deadlines.

It was here where he unearthed valuable insights on how to structure pre-trade data and documents, invent an intuitive and expedited onboarding process powered by patented AI in order to resolve many of the trading and posttrade issues, leading to his creation of his own fintech startup: Saphyre.

SAPHYRE

“ People want things that are cutting-edge but simple enough to address the complexities between different financial institutions and the accompanying regulations” STEPHEN ROCHE

PRESIDENT, SAPHYRE

56

August 2021

SAPHYRE

for an all-out focus on customer acquisition that emphasises its cutting-edge solutions.

Big-name partners like BlackRock help to solidify that reputation and the Roche brothers are convinced that their platform’s superior customer onboarding experience will ensure continuous uptake momentum.

Aware of the difficulty and expense of developing solutions comparable to Saphyre, Gabino suggests that introducing the company’s platform as a third-party

application is becoming an increasingly popular option. “Instead of spending millions to build something, it’s much easier for customers to ‘white label’ our technology, put their name on it, and then add ‘powered

by Saphyre.’” Doing so would enable even greater go-to-market agility for clients, a vital advantage as digital finance becomes evermore highly regulated. Positing that regulators act in broad motions because they cannot achieve a more intricate picture of data, Gabino claims that “Saphyre can change those behaviours by being forward-thinking and thereby saving institutions millions in fines.”

Asked to summarise what Saphyre’s primary goal for 2021/22 will be, Stephen emphatically replies, “Scale.” With the business growing and uptake increasing, the company is keen to shed the image that it’s only focused on pretrade. Instead, Saphyre will emphasise how diverse and complementary its solutions can be at a time when FSIs are searching for ways to upgrade their systems but unsure of how to do it. “For years, we’ve been saying that our model is sophisticated yet simple; people want things that are cutting-edge but simple enough to address the complexities between different financial institutions and the accompanying regulations.” It is Saphyre’s ability to marry realworld problem solving and interoperability that will ultimately produce its success, concludes Gabino. “A lot of firms are invested in pseudo monopolies that prevent data from being democratised because it's how they make money. Saphyre’s customers know that their data isn’t stuck on a single system. They retain ownership, they can port it, and they can demand innovation from their tech partners to keep data flowing end to end.”

aimagazine.com

57

MACHINE LEARNING

NLP: FRACKING for Data Natural language processing (NLP) is pushing our ability to mine data ever deeper, and humanising the sort of data we collect

WRITTEN BY: ELISE LEISE

58

August 2021

W

hen you take AI and focus it on human linguistics, the result is natural language processing (NLP). It is a branch of artificial intelligence that enables machines to understand the human language and its goal is to make sense of text and automatically perform tasks such as translations, spell check and subject classification. Basically, it enables computers to understand, interpret and manipulate human languages.

Its roots are in linguistics and speech recognition where it emerged to enable computers to literally process natural language and may use machine learning, as well as deep learning to effectively process speech and text datasets.

Despite the fact that NLP usually is used in many consumer applications, it also has significant implications for organisational IT. This is because of our use of email, chatbots, social media feeds, and documentation like contracts and claims forms. It categorises content, analyses sentiments, and summarises documents and can be broken down into other tasks based on the enterprise’s needs.

MACHINE LEARNING

aimagazine.com

59

Help us in the search for the Top 100 Leaders in Technology NOMINATE NOW

A BizClik Media Group Brand

NOMINATE

SHARE

CELEBRATE

Creating Digital Communities

MACHINE LEARNING

Natural language processing The study of NLP began in the 1950s with the first attempts of automated translation from Russian to English, laying the groundwork for the research. At around the same time, the Turing Test - or imitation game - was developed to determine if a machine was capable of behaving in the same way as a human.

“ If there is any racist or sexist bias in society, this could produce data reflecting this, which can exacerbate problems and is particularly troublesome in security, policing, health and employment” SHIVVY JERVIS

TRENDS FORECASTER AND CEO, FUTURESCAPE 248

Challenges of enterprise NLP Of course NLP applications come with the same risks of failure as any other AI deployments, including inflated expectations, unclear business cases and a lack of training data, which require different training sets, depending on the language processed and context.

For example, one set of training data may be needed for processing contracts and another set for solutions to payroll queries. Then there are language nuances, such as sarcasm, local dialects, intent and generating bespoke responses. aimagazine.com

61

MACHINE LEARNING

However, if a business is struggling to picture where NLP will fit into its work, it should be considered that the technology is much further reaching than the more recently developed smart assistants.

Everything from search, email spam filtering, online translation, grammar and spell checking, plus many other applications use NLP. Any kind of machine learning done involving natural language will involve some form of NLP. Shivvy Jervis, Trends Forecaster and CEO of Futurescape 248, touches on bias as one concern. “If there is any racist or sexist bias in society, this could produce data reflecting this, which can exacerbate problems and is particularly troublesome in security, policing, health and employment. However, handled correctly, it is an extremely useful tool for increasing productivity and security and enabling businesses to run efficiently and at lower cost”.

Trends sub More and more types of business are starting to see the benefits of NLP and although budgets have been hit hard by the pandemic, budgets for it are 10% higher compared to 2019. The latest trends point towards applications that require limited labeled data and simplified processes that make NLP accessible to everyone.

Transfer learning is a machine learning technique where a model is trained for one task and repurposed for a second that’s related to the first one. This means there is no need to build a model from scratch, which is expensive, time-consuming and requires large amounts of data, and NLP tasks can be completed faster using smaller amounts of labeled data.

62

August 2021

“ [Use cases for NLP] can be from automatically converting a scanned document into a text file, deciding what movie to recommend to you next, or deciding future education, finances or opportunities” DR GARFIELD BENJAMIN

POSTDOCTORAL RESEARCHER, SOLENT UNIVERSITY

MACHINE LEARNING

Another breakthrough has been the creation of machine learning models that create articles from scratch, using Generative Pre-trained Transformer 3 which can understand the context of words in a way that wasn’t possible previously.

Additionally now people don’t necessarily need the solid coding skills once required to work in this field. Lowcode and no-code tools have been around for some time, but are set to become commonplace, with SaaS companies democratising NLP and machine learning.

This would allow non-technical users to perform NLP tasks once only accessible to data scientists and developers.

Kiran Matty, Director of Product Management, Aerospike “The technology is being democratised so is no longer the ‘prized possession’ of a few chosen technology companies. We see data serve as the backbone for our customers of all sizes, spanning Oil & Gas, Automobile, Financial Services and advertising. The ability of these companies to transform raw data into insights and then decisions that drive remarkable financial and productivity outcomes never ceases to amaze us”

aimagazine.com

63

MACHINE LEARNING

“ With so much data at our disposal it is therefore crucial to understand it, monitor it and in some cases, censor it” 64

August 2021

MACHINE LEARNING

Also, until now, most advances have been focused on English, but big tech companies like Google and Facebook are now publishing pre-trained multilingual models which perform just as well or even better than monolingual ones. Open-source libraries are following in the footsteps of the two companies so it is expected there will be more NLP models emerging.

Dr Garfield Benjamin, a Postdoctoral Researcher at Solent University, says there are seemingly endless use cases available.

He says: “This can be from automatically converting a scanned document into a text file, deciding what movie to recommend to you next, or deciding future education, finances or opportunities. In every sector it seems you can find someone peddling an AI system to make decisions, across businesses, government, charities, education, health and the military the list goes on”.

What is still to come in terms of the technology?

NLP is rapidly evolving, with applications of its processing growing constantly. With so much data at our disposal it is therefore crucial to understand it, monitor it and in some cases, censor it, for example in the case of cyberbullying. Thanks to ready-to-use pre-trained models, along with low-code, no-code tools that are accessible to everyone, in the years to come, NLP is set to become even more widespread. Businesses in particular will continue to benefit from the technology, from improving their operations and customer satisfaction, to reducing costs and making better, informed decisions.

Now more than ever, businesses of all kinds need a helping hand and for this NLP could be just the ticket. aimagazine.com

65

66

August 2021

CPQi

BANK ON THE RIGHT OUTCOME AND FUTURE PROOF YOUR INSTITUTION WRITTEN BY: DAN BRIGHTMORE

PRODUCED BY: TOM VENTURO

aimagazine.com

67

CPQi

Creating an ecosystem of partners and solutions to drive the digital transformation of financial services

C

PQi is headquartered in Canada and operates across the Americas in Brazil, the USA, Chile, Argentina, Mexico, Colombia and Peru. Exclusively offering Fintech services managed by experienced banking-trained staff, CPQi partners with the likes of Calypso, Murex, Moody’s and Finastra offering solutions that substantially reduce costs for complicated major platforms.

Digital Transformation Digital transformation has swept across the banking industry. As the provision of financial services adapts to the developing digital demands of consumers, key trends have begun to emerge. Reacting to those trends, CPQi boasts the successful delivery of over 25 platform projects and numerous build projects with a keen focus on digital transformation. The company’s CEO, Terry Boyland, has positioned CPQi to support the technological evolutions reshaping how the world does business via four key pillars: Cloud, DevOps, Artificial Intelligence and an Omnichannel strategy.

Toronto business district, Canada

68

August 2021

Cloud Migration “What is digital transformation?” asks Boyland. “We have our own philosophy at CPQi where we look at it in four different ways… It begins with migration to the cloud, which allows us to do three things: increase agility, lower budgetary expenses and move away from the CapEx associated with legacy infrastructure.” Boyland notes cloud is a managed service in its own right, where the need for people to be involved is limited to a certain number – a plus during the pandemic. Scalability is key.

Using onsite equipment to protect sensitive

CPQi

aimagazine.com

69

CPQi

CPQI: driving banking forward

financial information can be costly. This is especially true when business and demand scale up. With the cloud, your platform scales automatically and can handle heavy workloads as and when required, enabling remote work environments and greater efficiency.

DevOps “The second path towards digital transformation for CPQi is the way in which we deliver new technologies,” explains Boyland. “The methods have changed; gone are the days of the waterfall when things were designed by business analysts, signed off in blood, went through coding – change controls were huge. They went through testing, and by the time you got to the end of a three-year program, some of the things you were delivering were no longer relevant.

Today, with the agile methods underpinning approaches like DevOps, these methods are more important where we're looking at what we can get out of the squad within a 70

August 2021

given time through user stories. And that's where our geographical spread comes into play because squads need to be awake at the same time to be highly effective. The second pillar is the introduction of Agile methods and, and the cultural change around the delivery of DevOps.” DevOps prioritises the speed of delivery without sacrificing quality. It’s a method requiring continuous development and

“ As business models evolve in this fastchanging world, we know that the ability to be agile is essential to success” TERRY BOYLAND CEO, CPQI

CPQi

TERRY BOYLAND

2007

TITLE: FOUNDER & CEO INDUSTRY: COMPUTER SOFTWARE

Year founded

LOCATION: CANADA

250+

Employees trained in both banking and technology

Terry is board level professional and one of the leading names in the Americas for the investment banking, Digital Transformation and technology industries. With over twenty five years of senior experience he is currently Chief Executive of the CPQi Group, the leading IBTech services provider for the Americas and was a co-founder of Cubelogic (bought by Openlink) as well as MDX Technology. A regular speaker and writer he has appeared in conference for The Economist, Financial Times, AITEC and the DTI. He has had articles published in The Banker and The FT along with other journals.

30+

Clients in over 11 countries across the Americas

Predictive Technologies “People talk about AI,” reasons Boyland.

“And when you think about what they're trying to do with it, a lot of organisations we've worked with are employing teams of data scientists simply to be able to find out what their data looks like. But we look at it differently. We say that predictive technologies, that's all they offer, predictions. If you look at your existing workflows and you find predictions in those workflows, you can change that prediction, whether it's done in Excel, for example, and replace it with an AI model and get a better-quality prediction that's far more reliable; then you are able to simply plug that into your existing workflow without re-engineering your entire business process and achieve some immediate benefit.”

The benefits AI can deliver include the automation of customer identification and authorisation tasks, the creation of

EXECUTIVE BIO

testing which offers many benefits for banking: continuously updated software deliveries; automation of many technical aspects and maintenance tasks that boost productivity; increased collaboration, and the creation of more time and opportunities for innovation.

Treasury & Capital Markets

Financial institutions of all sizes are tasked with powering local economies and providing best-in-class services for treasury and capital markets. This is no easy task as the complexities of the current economic climate have created more variables and uncertainty than ever before.

ENABLING ECONOMIES: Banks can further drive social inclusion by making finance available to more people. The funds allow economic agents – people – to invest in their ventures.

They help businesses launch, grow and diversify, strengthening local economies, and facilitating investment and opportunity in developing economies, thus enabling economic mobility. This is fundamental for robust and sustainable growth.

DRIVERS OF GROWTH: Banks clearly have the power to fuel growth and development. In fact, it has been argued that the smooth running of a country’s economic activities depends on an efficient banking system1, while others cite banks’ abilities to foster economic growth through funding productive projects as a prerequisite for economic growth.

LOCAL TO GLOBAL IMPACT: The impact of banking organizations can be seen as well on a larger scale as multilateral development banks (MDBs) – such as the World Bank or a regional bank such as the EBRD, or African Development Bank – have a hugely positive impact on public health, peace, security, climate change and digitization.

Learn more FACEBOOK LINKEDIN TWITTER YOUTUBE

THE FUTURE OF FINANCE IS OPEN

CPQi

“ CPQi has the privilege of partnering with Finastra – one of the top three Fintechs in the world today. We also partner with the other major platform delivery companies for trading risks” TERRY BOYLAND CEO, CPQI

74

August 2021

CPQi

responsive conversational interfaces (chatbots) for front office banking, and enhanced fraud and risk management delivered with AI’s more efficient analytics capabilities.

Omnichannel Strategy As the financial services industry increases its digital presence, accessibility has become a priority. “You must be able to execute on any channel that a client wants to deploy,” affirms Boyland. “Channels have enabled us to electronically rebuild the communities we've lost during the lockdown. Well-constructed channels allow us to bring on board new customers and clients who are looking for a connected experience.”

Mobile accessibility is crucial – online, in-person, or over-the-phone business must be utilised to deliver an omnichannel experience. “Channel strategies are key to our development,” says Boyland. “CPQi works across all of these from north to south in the Americas to deliver meaningful change.”

The greatest change in the delivery of financial services in a generation Boyland believes we are in the midst of the biggest change in the history of banking…

“Three major themes are beginning to emerge now within the financial markets,” he says. “Open banking may be far more prevalent in Europe than it is here in the Americas, but the philosophy behind it is huge… simply stating that I own my financial data and that I can direct that financial data to wherever I want without constraints from a bank or insurance company. That basically gives me the right to build my own financial experience.

“And with that, I can construct my own Neobank, which is the second element we're seeing today. Fintech institutions that don't have a banking licence can borrow one from a bank. By using modern technology, all of the delivery methods are available to attack particular markets.

There are Neobanks focusing on the gaming community, others on the 18-25-year-old

DID YOU KNOW...

FINASTRA Finastra builds and deploys innovative, next-generation technology on its open Fusion software architecture and cloud ecosystem. Its scale and geographical reach enable it to serve customers effectively, regardless of their size or location – from global financial institutions to community banks and credit unions. “We’ve worked with Finastra since the early days when its Kondor system was part of Thomson Reuters,” remembers CPQi’s CEO Terry Boyland. “Kondor has been merged into

and built on via a suite of financial services applications from Finastra deployed under the Fusion banner. We believe the Fusion fabric is a tremendous product enabling organisations to use modern development methods through the cloud to access some of their legacy platforms and their existing engines underneath. We’ve worked with Finastra on a number of applications, both in Latin America and in North America.

They're great to work with and deservedly viewed as one of the top three financial technology firms in the world.”

aimagazine.com

75

“ Many organisations use the expression DevSecOps; for us, that’s already a key feature of our approach to DevOps – it’s fully included in the services we offer” TERRY BOYLAND CEO, CPQI

76

August 2021

CPQI

Digital Transformation in Financial Markets

aimagazine.com

77

CPQi

SECURITY During the pandemic, and with the growing shift to remote working, the work from home/work from anywhere dynamic has seen cybersecurity tested by phishing, scamming and intrusion. “It’s a hot topic, that’s getting hotter,” agrees CPQi’s CEO Terry Boyland. “Many organisations use the expression DevSecOps; for us, that’s already a key feature of our approach to DevOps – it’s fully included in the services we offer. Organisations must have policies to protect themselves, to begin with. You need access to control policies, IT security policies, and your virus software must absolutely be up to date; your controls on access to remote systems from home, they've got to be tested on a regular basis, and penetration testing must be done regularly across your platforms…”

More recently, CPQi has found a particularly annoying strand of spoof emails from senior execs, asking people to take some time out and go and buy some Amazon or Apple gift cards. “You’d be surprised how often they cause problems,” notes Boyland.

“When you consider building your IT applications going forward, security

DID YOU KNOW...

must be the core element across the

78

financial markets. One breach can be huge – it's not just the data breaches; it’s the breach of the financial markets and the systemic risk that’s involved. We just encourage all organisations, as we do, to retain the services of a CISO. That CISO must be fully involved in the life cycle of development and part of your methodology for project delivery.”

August 2021

market. These innovations don't come with a huge elephant-sized legacy of systems and processes, and they don't need a banking licence to proceed; they just need a bank to sponsor them.”

The change in open banking has ushered in the third biggest change – the rise of cryptocurrencies. “Crypto’s are becoming more mainstream,” highlights Boyland. “The American government is talking about a parallel electronic currency to the dollar, and Japan is considering the same. This should provide an element of stability, but what's more interesting is crypto 2.0 – tokenisation and the ability to represent one of, or any single asset as an electronic token, and to trade that asset on a blockchain-type system, as if it were a cryptocurrency.”

Partnering for success “CPQi has the privilege of partnering with Finastra – one of the top three Fintechs in the world today. We also partner with the other major platform delivery companies for trading risks. Murex and Calypso with Openlink are now part of the biome,”

CPQi

reveals Boyland. “One of our key partners is Moody's Analytics, which is more on the risk for a loan side. And then, with the move towards digital transformation, we've got established partnerships with Salesforce and with ServiceNow. All of those elements bring together a package enabling us to deliver both on the traditional training and risk systems and the move towards digital transformation.”

Boyland notes that many of the company’s partners have been undergoing their own transformations and migrations to the cloud – providing the opportunity to support and learn from each other. “It's not just about being able to get a major platform on the cloud; it's about the way in which the delivery of technology is done afterwards – typically using microservices and continuous integration, continuous deployment and automated testing in the cloud. And when you've got a platform that wasn't built for that, it can be quite challenging…

We’ve worked with many of our partners to

“ Channels have enabled us to electronically rebuild the communities we've lost during the lockdown. Wellconstructed channels allow us to bring on board new customers and clients who are looking for a connected experience” TERRY BOYLAND CEO, CPQI

understand how the functionality can be built out using the data that's sitting in those platforms without cannibalising their existing organisation – migration both on the grid and then onto the pure cloud for our partners.

For them, it's a different challenge because they have their legacy platforms out there.

For us, it's an opportunity, in addition to our partner work, to work also more lightly and exclusively on specific microservice-based cloud developments which are native to the cloud from the beginning.”

Expansion Boyland believes the future is very bright for CPQi. “There’s no let-up to demand right now. We work across all of the major economies in the Americas and successfully navigate their tremendous cultural differences.

If we look at Brazil with its spirit of adventure and incredible optimism, there’s a high level of sophistication in their banking systems which, in some ways, are far more advanced than North America. The offshore centre that we have in the North-Eastern city of Fortaleza still serves that community there. And what we're beginning to see is that as demand continues to increase dramatically, organisations that previously wanted their work done in major centres like Rio de Janeiro or in Sao Paulo are now

CPQi

willing to accept that these things can be done remotely. And that's going to be great economically for the North East of Brazil, both in Fortaleza and other areas we're thinking about.”

“Here in the US, there’s also a tremendous pioneering spirit as well,” adds Boyland. “They led much of the early offshoring to Virginia, and we've been delighted to divert some of that down to Latin America. And as we talk about the delivery of agile methods and the need for squads to be awake, there's a very clear understanding that same time zone delivering is important for our clients.”

80

August 2021

“In Canada, it’s a little more risk-averse,” observes Boyland. “In order to support that, we recently set up a centre in Halifax, Nova Scotia. We've gone from zero to 20 staff in the space of six weeks. And we're looking to expand that to around a hundred staff by the end of the year. It’s been driven by the high level of demand and the available pool of resourcing now that the acceptance of remote working is on the rise.”

Leading the future of banking innovation across the Americas Boyland is full of praise for a “phenomenal team” leading the implementation of the

CPQi

systems making innovation waves in the Latin American banking industry. “CPQi developed the first training systems that were done for certain types of derivative instruments in Chile and, when Brazil moved to a new system of retail trading derivatives, CPQi was proud to work with the stock exchange in implementing the platform that would register those types of trades (the post-trade registration of derivatives). We've implemented platforms in Argentina, which, following a period of stagnation, are the first major change in the Argentinian banking sector for many years. Meanwhile, up here in North America, we're pleased to work with

organisations like Manulife in helping them change the way in which they're using their training risk systems – particularly with the Bank of Montreal. We're also excited to work with Moody's Analytics, Cargill and MetLife in helping them improve the quality of their delivery to their internal customer base.”

Both in terms of the technology and the banking changes, CPQi’s teams have been able to lead. Boyland is proud to see his team maintaining the agility to consistently break new ground for the delivery of financial services.

aimagazine.com

81

AI APPLICATIONS

The FUTURE of CLOUD and AI Six shifts in cloud computing and AI that will impact companies into the next decade WRITTEN BY: ELISE LEISE 82

August 2021

C

ompanies may not have adapted nearly so well to COVID if not for the fusion of cloud and AI. In 2021, these technologies worked in tandem to enable contact tracing, home deliveries, and remote medicine trials. In the first quarter of 2020, as compared to the first quarter of 2019, spending on cloud computing accelerated 37% to US$29bn—and for good reason. The cloud provides space and security to store huge amounts of enterprise data, while AI analyses, categorises, and predicts what’s yet to come about. Coupled together, the two technologies are a perfect complement.

Here are six major shifts in cloud computing and AI:

1 | Multicloud. Provides scaled storage to gather data. As long as the network consists of 2+ clouds, whether all-private, all-public, or a combination of the two, it falls under this category. According to Gartner, 81% of organisations now work with two or more public cloud providers. This is because multicloud solutions distribute resources, minimise potential downtime, and protect companies from vendor lock-in. Essentially, they provide great risk mitigation against cyberattacks and network failures.

• Google Cloud Platform: Runs Google Kubernetes Engine (GKE), a cloud-based service that automatically deploys, scales, and manages computer applications. aimagazine.com

83

Get reliable network coverage and security protection, fast.

A modern network must be able to respond easily, quickly and flexibly to the growing needs of today’s digital business. Must provide visibility & control of applications, users and devices on and off the network and Intelligently direct traffic across the WAN. Be scalable and automate the process to provide new innovative services.

Support IoT devices and utilize state-of-the-art technologies such as real-time analytics, ML and AI. And all these must be provided with maximum security and minimum cost.

This is the power that brings the integration of two cloud managed platforms, Cisco Meraki and Cisco Umbrella. This integration is binding together the best of breed in cloud-managed networking and Security.

cisco.com

cisco

CiscoSecure

CiscoSecure

AI APPLICATIONS

What is AWS?

2 | Everything-as-a-Service (XaaS). Relies on outside expertise. To use platforms (PaaS), desktops (DaaS), software (SaaS), and all the rest, clients can purchase XaaS licenses for long periods of time under annual or bi-annual subscriptions. By 2024, the global XaaS market is expected to exceed US$344bn. To understand why smart executives will pay year after year for such a service, remember that XaaS suppliers can provide tech support at scale. As CEOs pursue lean business operations, they’ve started to realise that they don’t always have to hire more people to build value. Instead, they can outsource technical tasks, rely on experts, and concentrate on the areas in which they excel.

• Amazon’s AWS: Offers 175+ services for databases, analytics, robotics, ML and AI, IoT, mobile, and augmented reality.

3 | Public, Private, and Hybrid Clouds.

Helps with predictive analytics. On one hand, public clouds are essentially shared

infrastructure, with resources owned and operated by third-party cloud providers.

They’re low-cost, reliable, and take no maintenance—at least, not from you. On the other hand, private clouds run on proprietary infrastructure, offer more control, and allow for further customisation. Imagine the difference like that between a bike-share and a custom-built mountain bike.

Then there’s hybrid—the combination of the two—which companies have started to adopt at scale. Hybrid clouds balance security and flexibility, fitting regulatory requirements without forcing you to give third parties all of your data. Most importantly, companies use hybrid clouds to manage short-term spikes in demand. Hulu, Uber, and Airbnb all run several operations on their private servers and pay on-demand for extra cloud service when the need strikes.

• Netflix: Manages bandwidth spikes when half the nation watches Stranger Things by paying for on-demand cloud storage. aimagazine.com

85

AI APPLICATIONS

4 | Edge Computing. Enables IoT devices to manage data on-premise. This tactic, computing near the source of the data, is useful any time that companies want to amp up security and minimise delay. Imagine that you’re talking to your Alexa. Using the traditional cloud, there’s a noticeable lag as she processes your voice commands, sends data to the cloud, and receives your answer back. But if Amazon starts to use AI chips to bring cloud capabilities to the device itself, we’ll see lower latency, better predictive maintenance, and fewer centralised security hacks.

• Mercedes: Ensures real-time authentication between vehicles and servers to prevent its cars from cybersecurity attacks.

5 | Data Lakes. AI helps make sense of collected data. Essentially centralised storage areas for all of your structured and unstructured data, data lakes hold information from IoT devices, websites, mobile apps, social media, and corporate accounts—you name it. Once you have all the data in one place, you can analyse or integrate different data sources with the help of AI and ML to generate usable insights. Some even incorporate real-time data to help make decisions.

• Amazon’s AWS: Provides a data lake solution that allows companies like Netflix, Zillow, NASDAQ, and Yelp to tag, search, transform, analyse, and share subsets of data.

6 | Data Democratisation. Anyone can use AI insights. Self-service analytics tools have made it possible for the average non-technical user to access digital info without having to involve IT. In an MIT Sloan Management Review study, 77% of individuals 86

August 2021

Cloud Technology: What Is It? On-demand access to computing resources such as applications, data storage, and development tools. Since all the infrastructure is hosted by a remote data centre or cloud service provider, you don’t have to install or maintain it. Instead, you pay a monthly subscription fee or pay per use. Its

benefits: you can scale capacity up or down, spread your applications closer to global users (see: Edge Computing), and outsource IT expertise. If you’ve ever used Google to answer a question, Netflix to stream a video, or Dropbox to store a file, congratulations: you’ve already used the cloud.

reported that they now have increased access to useful data. As this trend grows more common, creative, marketing, and production divisions will have better analytic and predictive models at their fingertips.

• Walmart: Cuts down the amount of time it takes to solve complex business questions via its Data Café—its real-time, fortypetabyte cloud.

In a Think 2021 programme, IBM Chairman and CEO Arvind Krishna discussed how the agility of hybrid cloud and the intelligence of AI can accelerate digitalisation. Like any good pair, AI and cloud technology balance each other’s strengths and weaknesses—yin and yang for a digital future. The company concluded: ‘Every industry can build a stronger digital foundation to reinvent the way they do business and engage in the world’.

Yet the world is shifting under our feet. To truly adapt, your company should develop cloud self-awareness, talk to people who know the field, and find honest sources of feedback. If you go about it the right way, the cloud and AI can balance scale and security; consolidate consumer data; and help make sense of corporate confusion. But there’s no time to wait. As futurist, strategic technology advisor, and Forbes senior contributor Bernard Marr wrote: ‘Cloud computing will transform society’. aimagazine.com

87

DATA & ANALYTICS

88

August 2021

DATA & ANALYTICS

Following the

DATA SCIENCE Why is data science so important, what are the emerging trends and where is its future?

WRITTEN BY: LAURA BERRILL

D

ata science, or data-driven science, basically enables businesses to make better decisions. It allows them to find the leading cause of any problems by asking the right questions, perform explanatory studies on the data and model it using algorithms. The results are then communicated and visualised via whatever means necessary, for example graphs or dashboards.

“ Every modern business has the ambition of being data led, meaning vast amounts of data is being accumulated” RAJASEKAR SUKUMAR

VP EUROPE, PERSISTENT

It’s directly because companies across all sectors and industries are sitting on an ever-growing treasure trove of data that modern technology has enabled the creation and storage of increasing amounts of information and as a result, its volumes have exploded. It is estimated 90 per cent of the data in the world was created just in the past two years. However, this data is still often sitting in databases and lakes, mostly untouched. That’s where data science comes in. Data is the bedrock of innovation, but its actual value comes from the information data scientists can glean from it and then act upon in a business’ best interests and objectives.

A high flying example In practice, data science is already helping the airline industry, as one example, by helping predict disruptions in travel in order to alleviate the pain for both airlines and passengers. With the help of data science, airlines can optimise their operations in many ways. These include planning routes and deciding whether to schedule direct or connecting flights; building predictive analytics models to forecast flight delays; offer personalised promotional offers, based on customer booking patterns and decide which class of planes to purchase for better overall performance.

There are some main technical concepts to explore before any move into data aimagazine.com

89

DATA & ANALYTICS

ACCESS THE WORLD’S LEADING

TECHNOLOGY COMMUNITY

September 14th - 16th SPONSORSHIP OPPORTUNITIES AVAILABLE

CONTACT US NOW

BRAND AWARENESS | THOUGHT LEADERSHIP | NETWORKING | LEAD GENERATION

90

A BizClik Media Group Brand August 2021

Creating Digital Communities

DATA & ANALYTICS

“With data sets going back years, it’s only possible to really leverage all of this data with data science and machine learning as a helping hand” DAVID SEMACH

PARTNER AND EMEA HEAD OF AI AND AUTOMATION, INFOSYS

science and the employment of associated data scientists. Machine learning is the backbone of data science, so data scientists need to have a solid grasp of this technology, along with a basic knowledge of statistics.

Modeling is also part of machine learning and involves identifying which algorithm is the most suitable to solve a given problem and then to train these models to do this going forward. As well as this, some level of programming is required, the most common of which are Python and R. Python because they are easy to learn and support multiple libraries for both data science and machine learning. Finally, data scientists need to

understand how databases work, how they are managed and how the data is extracted from them.

Growing trends Rajasekar Sukumar, VP Europe at digital engineering and enterprise modernisation partner, Persistent, says one growing trend is edge analytics and he predicts the technology will make its way into the mainstream within the next two to three years.

“Every modern business has the ambition of being data led, meaning vast amounts of data is being accumulated. All of this aimagazine.com

91

DATA & ANALYTICS

has to be managed and processed into a suitable format to get any useful insights and, as it stands, the whole model can become inefficient and costly. With edge analytics, organisations can let the model lead the process, rather than analysing the data in order to build the model. With the processing happening at the edge, where data is created and consumed, this can bring huge efficiency gains and reduce costs in the long run”. Sukumar gives Tesla as an example. “The car is collecting vast amounts of real-time data, so the most efficient way to analyse this is to build a model and enable all the actions to happen there directly, rather than retrieving the data, sorting it and then pushing it back to the car.”

Sukumar goes on to suggest some of the fastest growing data science trends for this year and beyond. One is Deepfake video and audio, which use AI to manipulate or create content to represent someone else, with images or audio modified to someone else’s likeness or sound. He says there is huge scope for this to be used maliciously via what is known as ‘voice phishing,’ which could be used against individuals such as politicians. Because of this threat, governments are starting to look at legislation and social media regulation to defend against this, along with technology that can identify deepfake videos.

Additionally, there is increased demand for end-to-end AI solutions. For example, Dataiku, a start up of which Google bought a stake. It helps enterprise customers clean their large data sets and build machine learning models. This way, big companies like General Electric and Unilever can gain deep learning and valuable insights from their massive amounts of data, as well as automate important data management tasks.

92

August 2021

Data Science The term “data science” has been traced back to 1974, when Peter Naur proposed it as an alternative name for computer science.

In 1996, the International Federation of Classification Societies became the first conference to specifically feature data science as a topic.

DATA & ANALYTICS

“ Mature organisations are starting to change their strategies and tactics in how they organise their data people and how they build their data science platforms” ANDREW MORGAN

DIRECTOR OF DATA, 6POINT6

So how is data science changing and why?

David Semach, Partner and EMEA Head of AI and Automation at Infosys Consulting says he believes there will also be a huge move towards using the combination of AI and data science to make predictions for the future.

“This will enable business leaders to make informed decisions based on much more data than they ever have before. In Fortune 500 companies, when senior executives make decisions, it has typically been based on past

events within the company, plus some manual forecasting on what might happen next. With data sets going back years, it’s only possible to really leverage all of this data with data science and machine learning as a helping hand.”

There are other changes afoot in the industry, according to Andrew Morgan, Director of Data at 6point6 and author of ‘Mastering Spark for Data Science, which he thinks indicates we are moving into a new era for the science.

“Mature organisations are starting to change their strategies and tactics in how they organise their data people and how they build their data science platforms. The first new trend we see is the merger of data science teams with general data analytics teams, into larger organisations focussed on Data Enablement.

The second is a move to investing in technologies that incorporate data engineering best practices to help Data Scientists meet a rising tide of AI industrialisation expectations.

The third is that data scientist teams are migrating to the cloud, and to move this along, they are engaging third party suppliers, often for the first time,” he says. aimagazine.com

93

DATA & ANALYTICS

“Whether a business is tech-savvy or not, data science culture must be embraced so analytics becomes an everyday part of business operations” JAMES MCELHONE

PARTNER IN CONSULTING, EYE

94

August 2021

DATA & ANALYTICS

Douggie Melville-Clark, Head of Data Science at Duco “With data becoming such an integral part of every industry, education needs to follow. A shortage of analytic skills will set industries back, with employees unable to perform the statistical tests and experiments large data sets need”

The future of data science Others in the know feel there will be an increasing emphasis on software engineering and more of an embrace of software craftsmanship. Matt Shearer, COP at Data Language believes harnessing data science at scale means it is about 90% engineering and 10% the machine learning aspect. He adds: “What’s positive is that we are seeing this understanding translate into action.”

There is also the consideration of pairing human experience and judgement with technology, extracting the maximum value from data, while mining the potential for human leadership, creativity and empathy. Helena Schwenk, VP at Exasol’s Chief Data and Analytics Office says, “Collaborative intelligence will help optimise and enhance aspects of customer communications in high value areas such as fraud detection, customer service and issue handling. This is where AI will complement and augment human capabilities, not replace them.”

And James McElhone, Partner in Consulting at EY, says despite all the advances in technology, whether ingrained or still part of blue sky thinking, data science cannot be used in isolation. He thinks it must be part of an overall business strategy going forward and not determined by the capabilities of an organisation itself.

“Whether a business is tech-savvy or not, data science culture must be embraced so analytics becomes an everyday part of business operations. AI also has the potential to miss opportunities and that is where human-inthe-loop comes to the fore. Organisations are better placed if they combine the benefits of an infusion strategy, embedding analytics and AI into the core of the business process and supporting an increased skill set in the workplace. The future is neither one nor the other, it is a combination of the best of both worlds. aimagazine.com

95

TOP 10

96

August 2021

TOP 10

Artificial Intelligence companies to watch this year

Analysts IDC forecast Artificial Intelligence’s fiveyear Compound Annual Growth Rate will be 18.4% in 2021, with revenues reaching $37.9 billion by 2024

A

rtificial Intelligence is becoming a growing force in business, with entire industries being reshaped by the technology. Additionally, AI companies attract massive investment from venture capitalist firms and the tech giants such as Microsoft and Google recognise the potential for further growth in corporate and personal use. AI research is growing quickly in quantity and complexity, as

are AI job openings across a multitude of industries.

Companies are leveraging cloud and edge computing, with machine learning leading the pack at the moment. However other technology categories are in the running, from predictive analytics to business intelligence and data warehouse tools, to deep learning. As such, today’s top AI companies are leaders in what is still an emerging technology.

aimagazine.com

97

TOP 10

10 Ascent

Ascent was founded in Tokyo in 2016 and delivers scalable and dynamic software for machine autonomy. It also works in research and development in Artificial Intelligence and simulation. It’s first advanced robotics grasping product is a software-based AI system that enables object manipulation in any type of environment. It also develops autonomous systems for mobility applications, using a combination of advanced simulation and real-life data.

Ascent's first robotics grasping product is a softwarebased AI system 98

August 2021

09 Osaro

Osaro focuses on AI for industrial-scale automation, building the technology for factories, warehouses and logistics systems. It was founded to build the technology and products that make robots smarter and to lead in practical revenue-generating applications.

The company brings together advanced machine learning for visual perception and powerful control software for object manipulation, its software also leverages flexible and scalable intelligence that adapts to customer-specific use cases.

08

Automation Anywhere A vendor in the small, but growing Robotic process Automation market, the company actually created the term RPA. It claims to be the number one RPA platform which is easy to use and scales three times faster, at one fifth of the cost of legacy platforms.

It employs Automation 360, the world’s only cloud-native, webbased platform for end to end automation, with double the amount of automated processes.

07

Salesforce Salesforce is a CRM and SaaS, cloud and enterprise computing provider and is now also acquiring AI companies and has its own AI service, Salesforce Einstein. It is made up from a layer of intelligence within its Lightning Platform, enabling admins and developers to build smarter apps and customise AI for their business. Its most recent project utilises machine learning to help workers more proficiently perform tasks.

aimagazine.com

99

Help us in the search for the Top 100 Leaders in Technology NOMINATE NOW

A BizClik Media Group Brand

NOMINATE

SHARE

CELEBRATE

Creating Digital Communities

TOP 10

05 Alibaba

06 Nvidia

Nvidia is one of the oldest established Artificial Intelligence companies, established in 1993 and still plays a significant role today. Its graphic processors (GPUs) are extremely important to today’s AI and Machine Learning industry. The Delawarebased firm contributes to medical care, higher education, retail and robotics technology.

It is particularly focused on incorporating AI into vehicles, in manufacturing and in autonomous driving. Its GPU deep learning is available on services from Amazon, Google, IBM and Microsoft, among others.

Asia’s powerful cloud computing platform offers customers a refined Machine Learning Platform for AI via a visual interface. This is so organisations can relocate different segments into a canvas to gather their AI functionality. Also incorporated into the platform are scores of algorithm components that can deal with several tasks and enable users to utilise pre-built solutions. The company was founded in 2009 and in 2017, the company became the official cloud services partner of the Olympic Games.

TOP 10

04 Amazon

Seen as the ‘pioneer’ in cloud computing, Amazon Web Services (AWS) offers both consumer and business-centric AI products and services and a large number of its AI services are created on consumer products. AWS says it offers the broadest and deepest set of machine learning services and supporting cloud infrastructure, claiming to be helping more than 100,000 customers accelerate their machine learning journey through putting the technology into the hands of the developer, data scientist and practitioners.

102

August 2021

03

DataRobot Lesser known, DataRobot’s artificial intelligence platform automates the whole end-to-end data science. It gives the results of automated machine learning, MLOps and automated time-series.

Different models are provided with DataRobot like Managed AI Cloud, On-Premise AL Cluster, Private AI Cloud and Hybrid AI Cloud. Its enterprise AI platform accelerates and democratises data science by automating the endto-end journey from data to value, enabling users to deploy trusted AI applications at scale within an organisation.

Apple Apple is similar to Google in that it has also been occupied with procuring AI start-ups and considers AI technology as a crucial element of its future. Just three years ago the organisation hired computer scientist, John Giannandrea, as its head of the AI and Machine Learning office, back from Google. AI and machine learning within the company is done by managing the development of

“ I knew there was much machine learning Apple should do, but it was not being done.

That has changed dramatically”

different products, for example, Siri and its new Create ML tool, which MacOS and iOS developers can use to make productive and simple training courses for all their apps.

Whereas Google has a reputation for participating in, and in some cases leading the Ai research community, Apple used to do most of its work behind closed doors. This has changed in more recent years.

02

aimagazine.com

103

THE ULTIMATE TECHNOLOGY, AI & CYBER EVENT SEPTEMBER

14th - 16th STREAMED LIVE FROM TOBACCO DOCK LONDON A BizClik Media Group Brand

Confirmed Speakers Include: Danny Attias

Chief Digital & Information Officer Anthony Nolan

Marcell Vollmer

Chief Digital Officer Boston Consultant Group

Kate Maxwell

Chief Technology Officer Microsoft

Renata Spinks

CISO United States Marine Corps

Scott Petty

Chief Technology Officer Vodafone

EARLY BIRD TICKETS HERE

Creating Digital Communities

TOP 10

Why Google's DeepMind Is The Future of AI

“ At Google, we believe that artificial intelligence can provide new ways of approaching problems and meaningfully improve people’s lives”

106

August 2021

Google Hardly a shocker, Google is on an enormous AI acquisition binge, having procured a bunch of AI startups in recent times. Along with utilising Ai to improve its services, Google Cloud sells various AI and machine learning services and has a cutting- edge software project in TensorFlow, as well as its own Tensor AI chip project. The global tech giant believes AI is making

it easier for people to do everyday things and breaks down barriers in life, as well as providing new ways of looking at existing problems, from rethinking healthcare to advancing scientific discovery.

One of their main areas is AI for social good, a programme focusing Google’s AI expertise on solving humanitarian and environmental challenges.

aimagazine.com

107

Enabling educators.

Empowering students.

Explore how we accelerate student discovery, learning and innovation with our Digital Education 3D Experience.

EXPLORE THE 3D EXPERIENCE