Computer psychos

Page 1


Technology In Medicine By Eden Hanlon

16

Technology in modern medical procedures and operations.

The Technicalities: An Analysis of Algorithms

8

By Nir Elber

An overview of machine learning algorithms.

Editor Letters

4 2 Computer Psychos

Writer’s thoughts about the subjects in the magazine.

The Table


AI In Business and Markets By Deven Sterzing

22

AI in modern day bussiness and market.

Siri Live My Life By Timothy Edison

28

The use of AI in day to day activities.

Video Game AI

Of 34 Contents

By Ikthan Marroquin

AI in video games, it’s characteristics and features.

3 Computer Psychos


Letters from the Editors

The writing group in its natrual habitat, awkwardly standing and sitting around and smiling. Top, from left to right: Ikthan, Deven. Bottom, from left to right: Timothy, Nir, Eden.

4 Computer Psychos


I was born in Austin, Texas, but I then quickly moved up north to Pennsylvania and New York. While living up North, I grew to love technology and AI. I started to code, and became interested in AI, and how it works. I started to think about how AI could be implemented in a business. As I got older, I began to see how AI can and is used in marketing as well.

Deven S.

Fact Check Editor

When I moved back to Texas, my mom started working for an AI company called Cognitive Scale. Cognitive Scale develops AI software for other companies to use, as well as creating programs that do not only think for themselves, but augment the worker. Through her company, I was able to expand my knowledge of AI and how it is used in business.

I was also able to learn how marketing is used to present me with targeted advertisements that I am more likely to enjoy. I began to see how I was receiving targeted advertisements every time I went online. I love learning about AI and how it is implemented in business and marketing, so I was inspired to write a feature article about business and marketing.

Video games have always been part of my life, giving me countless hours of fun and wonder. They provided the opportunity to escape the real world and venture into one of fantasy, adventure, horror, you choose. This is why I would like to design or program video games for a living, because they have been more than just a mean to have fun in my free time, they have created awesome memories as well as connecting me with my now best friends.

Ikthan M.

Features Editor

I believe that the implementation of artificial intelligence into the video game world is a huge asset for the industry that will allow a more immersive experience for the users as well as a step up in the video game evolution. Upgrading the use of artificial intelligence will give NPCs the ability to be more human-like which will provide a smarter and more realistic gameplay.

The one big obstacle I faced when designing and writing the article was the lack of people that specialize in the subject within a reasonable range. This limited my sources of information and kept me from adding more professional commentaries. Finally trying to find usable images or the article was close to impossible, due to most companies copyrighting their game images and logos.

5 Computer Psychos


Letters from the Editors Programming has been passions for 3 years know, and no matter what happened, I always stuck with it. I originally began with work in HTML and CSS and have now tested myself in many fields both technical and mechanical. I wouldn’t consider myself particularly good at programming, but the world of tech has always interested me.

Timothy E. Layout Editor

I chose “AI in everyday life” because it would probably be the most applicable to our readers. Technological innovations are incredible and most people just a bat an eye whenever their phone does a command. The real beauty I have found is how these applications interact and overall create a friendly interface that anyone can understand.

As much as I already new about AI, I learned that I really controls almost every aspect of our everyday lives. Many people simply bat an eye at its advancement of this technology as is it has so effectively weaved its way into society. Once considered a thing of the future, it is now playing a major part in our everyday lives. Also, it was interesting to learn how AI works and its modern applications throughout the other articles in this magazine.

I’ve always liked programming. Something about the telling a computer what to do and then having the computer actually perform it has always intrigued me. So when I was introduced to machine learning, I didn’t even have to properly tell the computer what to do. Just tell the computer the problem and some example solutions, and suddenly the computer figures out how to solve the original problem. It’s pretty cool to watch it happen.

Nir E.

Copy and Content Editor

The problem with learning this topic is that one can’t go very far without seeing some calculus. And while the truly technically inclined can go off and learn the prerequisite theory, many people just don’t care enough. So, I wrote “The Technicalities” to aim to those people. The article was intended to be a friendly calculus-free introduction to machine learning. Just a few words, some algorithms, and hopefully a better understanding of the topic afterwards.

Of course, the difficulty of talking about something that involves calculus without ever mentioning calculus is that, well, I couldn’t mention calculus. So attempting such a task was helpful for my ongoing trials of explaining ideas in simple words. As hard as it was, at least it was a good experience.

6 Computer Psychos


Photography has been my hobby for several years now, and while this has nothing to do with the article I wrote, this was why I chose to select the magazine topic box as “Technology” when Ms. G put the groups together. While in school, the only subject I really truly enjoy is math. I’m sure this will change when I get to calculus, I’ve heard it’s less than desirable, but until then math is rather fun. Outside of school, besides photography, I do horse riding and martial arts. I’ve been riding horses for 8 years, and doing martial arts for 10 years. I find how technology affects people an interesting topic because of how technology has changed and evolved over time. I wrote about medical technology and how it is affecting Photo and doctors because doctors have a very important job, and I wanted to learn more about how Graphics Editor newer technology is affecting their work positively and negatively. From writing this article and working with this magazine, outside of the information from each of the magazines articles, I learned more about how time management is very important, and about the difference between writing an english paper and writing something for a magazine.

Eden H.

“The advance of technology is based on making it fit in so that you don’t really even notice it, so it’s part of everyday life.” -Bill Gates 7 Computer Psychos


Inspired by Black Hat from xkcd. 8 Computer Psychos


The Technicalities An Analysis of Algorithms By Nir Elber

T

his is Jimmy. Say hi to Jimmy. He doesn’t like reading poetry, so he has decided to enlist his computer to help him with his English homework. Specifically, he wants his computer to tell him what the poem is about without him needing to read it. But while the task seemed easy at the start, he has found himself quite stuck. Jimmy has stumbled upon the vast field of machine learning, and vast it is. According to Dr. Alexander Huth, an assistant professor at the University of Texas, machine learning is, in a nutshell, “building algorithms that can learn from data to solve problems.”

“It’s building algorithms that can learn from data to solve problems.” -Dr. Huth In this case, Jimmy is trying to build an algorithm to solve the problem of analyzing the meaning of poetry. He has tons of examples of poems and meanings, so all he really needs to do is teach his computer to imitate the results of his examples. Eventually, Jimmy will end up building a black box (a complicated program that won’t really be explained in detail) which takes in a poem, a string of words, and spits out some simpler, easier to understand words. For example, Jimmy’s first black box took in the phrase “clever as a fox” and spat out “smile,” which wasn’t exactly what Jimmy wanted. Jimmy can make his black box better in a few ways.

Huth is an assistant professor at UT. (www.cs.utexas.edu)

First and foremost, he’s going to have to know just how bad his black box is at predicting. That was pretty bad, but Jimmy can quantify just how bad with something called a loss function. For the case above, Jimmy can have some pretty simple loss functions: If the desired output was “smart,” we increase the loss by one for each incorrect letter. While simplistic, this loss function will do for now. The first two letters were right, but the last three were incorrect, so the loss was three in this case. Now, Jimmy has a few options of trying to minimize how badly his black box performs; that is, minimize his loss function.

9 Computer Psychos


- Genetic Algorithms -

The simplest and thus first of Jimmy’s options is a genetic algorithm. It’s rather ugly and inelegant, but for many purposes, it sort-of gets the job done. It’s kind of like comic sans. The actual process behind a genetic algorithm is intended to mirror evolution. First, Jimmy chooses a bunch of “locations” or “species.” This essentially means that he makes a whole bunch of black boxes. For the sake of limiting confusion, let’s say that he has red, blue, and green black box. Secondly, Jimmy has to decide which of his black boxes are the best. He’ll try a few examples; maybe he’ll input “cut the cheese” and expect the computer to give him “fart.”The red black box outputs “happy,” the blue black box outputs “family,” and the green black box outputs “part.” In terms of our letterwise loss function, the green and red black boxes did better than the blue black box. Third, Jimmy eliminates the worst-performing black box and replaces it with an average of the best ones. So, the green box goes and gets replaced by a purple black box!

Red boxes plus blue boxes make purple boxes! In practice, the two boxes are actually “averaged.”

Finally, Jimmy has to slightly alter each of his black boxes. This imitates the mutation aspect of evolution, which essentially keeps all of his boxes from becoming the same. Theoretically, because he replaced the worst with the average of the best, his black boxes should be better at their job now, and repeating the above process would make the boxes better and better.

Way too many hours later

Genetic algorithms are really slow. Hours of no change to be seen.

10 Computer Psychos

However, Jimmy soon begins to implement this algorithm, and he finds that his algorithm is unbearably slow. Upon looking online, Jimmy finds the problem: according to Dr. Philipp Krähenbühl, also an assistant professor at the University of Texas, genetic algorithms “don’t scale well as the model gets larger.” For example, Jimmy may want up to 50 differently colored black boxes to test and reproduce simultaneously. If he was really dedicated, he may use up to a few hundred! This is, obviously, extremely computationally intensive, so as Jimmy begins to run his genetic algorithm, he asks around to see if this is something that he really wants to do. According to Dr. Huth, he would use a genetic algorithm “very rarely” for the simple reason explained above: It takes way too long, and much better options exist. The main issue wit h genetic algorithms is the “mutation” step; specifically, the algorithmmodifies the back boxes randomly, which means that it may take ages before it picks upon a crucial change. So, Jimmy could have a significant improvement if his algorithm knew which way to go.


- Hill-Climbing -

The hill-climbing algorithm is step up. This time, Jimmy only works with one black box, which is already a great improvement. However, instead of evaluating one example once, Jimmy takes his black box, slightly alters one thing, e.g. makes the black box slightly redder, and sees how it does. Then, Jimmy does the opposite change, e.g. makes the black box slightly more cyan, and sees how this one does. Now, Jimmy can compare which black box does better, and he simply replaces the old black box with the better one. (If the original box was better than either, he keeps the original.) Even though Jimmy still has to look at every single variable, this algorithm is much better because only uses one black box and he quickly knows how to alter it. For this reason, hill-climbing is like Times New Roman: It works, but Jimmy could still do better. The issue with this algorithm is that for each variable, the black box must be tested twice. This can cause the same issue that we had before: It won’t scale well as the model gets larger because there are tons and tons of variables. To fix this, enter our next algorithm.

The hill-climbing algorithm is a step up.

11 Computer Psychos


- Gradient Descent -

“It’s all about finding the right variables.” -Dr. Huth

The gradient descent algorithm will be the last and most complex algorithm for this article. Excuse the handwaviness, but to explain gradient descent first Jimmy needs to consider what his black box is really doing: It takes in an input (a poem) and spits out an output (some words). His black box is essentially a mathematical function. His idea of “loss” is another function that takes some inputs (the black box and the desired output) and an output (an integer). So as said in the beginning, Jimmy is trying to minimize this loss function, but he needs to understand his loss function as a mathematical function.

Sometimes, functions have properties, like some functions are lines, and others are curvy. However, for this algorithm, the loss function needs to look a particular way; the keyword is “convex” or “differentiable.” For the sake of this explanation, consider this property to mean “smooth.” In the case that the loss function does have this nice property, Jimmy can feed in an example to his black box, and his black box will can give the output. Then, gradient descent can look at what the loss function says, and then it sort-of squints at it pretty hard, and then it sort-of squints at the black box pretty hard, and then it knows exactly how to change each variable. (If you’re interested in further details because this sounds too vague, read the next paragraph. If not, you may skip it. You won’t hurt my feelings.) To be more specific with this, the gradient descent more or less looks at a graph of the loss function. Being “smooth” in this case simply means that the graph doesn’t have any jagged corners or similar. If the given function is “smooth,” then the algorithm can take the input and the output and look around in the graph. Upon looking around, the algorithm can find which way it needs to go to make the loss function smaller, in that instant. See the following article (the ASF) for a more visual explanation.

“Never.” -Dr. Meller

This curve is “smooth.” So pretty. So nice.

“Ok, but what happens if the function isn’t convex, such as with poetry?” asks Jimmy. Well, not all hope is lost. Dr. Huth said that sometimes “you can tweak things about the problem to force things to look more convex,” which allows us to use gradient descent. Because this is possible so much of the time, actual researchers or similar almost always use gradient descent over genetic algorithms. When asked, three professors Dr. Krähenbühl, Dr. Huth, and Dr. Meller (from the University of Cincinnati) said that they would use genetic algorithms over gradient descent “rarely,”“very rarely,” and “never.” After hearing all of this, Jimmy begins to program in gradient descent and begins to teach it to imitate his poems and analyses. His English homework will be ready to be finished in no time.

12 Computer Psychos


-~-

While Jimmy waits, he considers how amazing it is that his machine is learning to imitate this. The problem was complex to say the least; heck, it could probably do better than some of his classmates! Marveling at his black box as it learns to do his English homework, he wonders if one day machines will be able to do all jobs. Dr. Krähenbühl agrees with him: “The end of machine learning is when it takes over every aspect of our lives.” But for now, Jimmy waits. That evening, he thinks that his computer should be done being taught. Jimmy inputs “Machines” by Kyle Kind: Left alone for far too long, The machines will take their rise. Rhythmic steps marching along. The fire will dance across the skies. Planes will fall, missiles will fly, This all seems like a dream. Nobody is safe, you cannot hide, In a world run by machines. And of course, his output is: Cats, cats, cats. The internet is made of cats.

Above cat inspired by minutephysics. Isn’t the cat cute? 13 Computer Psychos


Some Examples of Maximization In the feature article, “loss minimization” is mentioned. This is the idea that a “loss” function (describing how bad our network is) can be minimized. For a visual understanding, we’ll consider a “reverse loss” function describing how good our network is. So now, we will try to maximize a “reverse loss” function. Below we go through an example of a few ways to do this “reverse loss” maximization:

Example “Reverse loss” Function

Goal: Maximize! Algorithm: Genetic Algorithm

1

Pick a set of random staring positions.

2

Remove the worst and replace it with the “child” (average) of the best two.

3

“Mutate” (move) everything slightly.

4

On average, we’re closer to the top of the hill now! Repeat steps 2 and 3!

14 Computer Psychos


Algorithm: Hill-Climbing Algorithm

1

Pick a random staring position.

2

Move it slightly in both directions. Choose the better of the two.

3

On average, we’re closer to the top of the hill now! Repeat step 2!

1

Algorithm: Gradient Descent

Pick a random staring position. (See graphic above.)

Follow the slope of the current position to go up: left or right.

2 3

On average, we’re closer to the top of the hill now! Repeat step 2!

Sources: “A Genetic Algorithm White Paper” (Pat Langley, John H Gennari, Wayne Iba) “Hill-Climbing Theories of Learning” (Salvatore R. Mangano) Neural Networks and Deep Learning (Michael Nielsen)

15 Computer Psychos


Technology In Medicine By Eden Hanlon

“What do you want to be when you grow up, kids?” ‘I want to be an astronaut!’ ‘I want to be a princess!’ ‘I want to be a superstar!’ ‘I want to be a doctor!’ These are seemingly ludicrous answers; like princess and superhero, but some children who say, “Doctor!” may soon become one. And they will constantly be surrounded by medical technology and be faced with the question of whether or not all these advancements are completely beneficial.

“Sometimes technology doesn’t benefit doctors as much as it benefits others” Technology has become an integral part of the field of medicine, but the tools that doctors are using is mostly being developed by engineers. Even though sometimes technology doesn’t benefit some doctors as much as others, it has been highly influential and helpful. Technology affects doctors with every new technological advance and change. When new technology is put into use, doctors must adapt to it. Most of the time, this new technology is beneficial to the medical field, but there are occasional drawbacks. For example, it could be excessively time-consuming, it could take away from the significant doctor-patient relationship, or it could sacrifice convenience. This seems to be an ongoing struggle in the medical field: balancing efficiency, effectiveness, and convenience for all parties involved, particularly the doctors and patients involved.

16 Computer Psychos

This is a headshot of Dr. Donna Ruesink in Austin Texas. (www.ruesinkdds.com)

“Technology has made my job more efficient in some ways and not in others,” says Donna Ruesink, a dentist for 31 years. According to Dr. Ruesink, machinery like X-Rays and improved lighting create better visibility, but other aspects take excessive time. Dr. Ruesink says that her job is more time-consuming than it should be: “Dealing with computer issues in general takes a lot of time. Training and maintenance of computers and programs require time and patience.” These things could greatly impact her work and cause stress due to lack of time left in the day. These are negative technological aspects of

“When I was in training, the majority of the images were hard copies” -Yip her work, but her profession in the dental field of course has benefits, “bonding agents and materials continue to be stronger and more esthetic.” Scientists are studying the regeneration of the tooth in laboratories, and engineers are making many new advancements, and for radiologists, this has been substantially beneficial. Radiologist Dr. Virginia


This image depicts whats is most likely a dentist’s chair taken in 2016. (www.pexels.com)

Yip wrote, “there are many changes, and constant advances made in imaging in the last decades.” Because the age of computers is developing exponentially, the majority of corporations, businesses, and hospitals are becoming more reliant on computers. For example, the way imaging used to be done is no longer effective, and so it has transitioned to a more efficient digital age. Yip wrote, “When I was in training, the majority of the images were hard copies. Nowadays, the images are in digital format.” This adjustment for radiologists from hard copies to digital ones took time but was well worth it. This is one of many examples. These advances have made her job, centered around recording images, easier and have allowed for higher-quality images to be captured. Doctors treat people, and therefore need to keep good records if anything ever went wrong in the event of a malpractice lawsuit. A malpractice lawsuit can occur when a patient sues a doctor or hospital for doing their job incorrectly or just not to the best of their ability. Doctors used to have a scribe that would write a report for them and have the doctor sign for the accuracy, but now most doctors use digital recording technology. Dr. Flora Edison, who has experience with a human scribe in the room along with a patient says, “the scribe is not a machine, it’s a person right there who’s typing out your conversation and it’s kinda distracting for the patient.”

By transitioning to digital recording, it has fixed Dr. Edison’s problem and has allowed for a better doctor-patient relationship. For example, when doctors are not with a patient, they or one of their nurses are filling out charts or documentation, which may take a lot of time. Dr. Joseph Brown, a former pediatrician and anesthesiologist, says, “Electronic health records document a lot more info than was typically recorded on paper health records. This data entry takes time, and someone has to do it.” When digital paperwork takes more This depicts an X-ray of a human head taken in 2011. (commons.wikimedia.org)

17 Computer Psychos


This image depicts a denstist table photographed in 2016. (www.pexels.com)

time, this means that, again, less patients may be seen in a day. New technology requires adaptation, which is part of a doctor’s job, but that doesn’t mean it’s easy. For Edison, certain adaptations have been harder than others: “It really has been difficult because when I started practicing medicine, we didn’t have computers, but now everything is electronic. All your data is on the computer.” However, because everything is on the computer, doctors must often look at a screen instead of the patient. While technology has negatively affected Edison’s job in some ways, it has also provided positive effects; she said, “before it was just your physical exam skills and a few basic tests but This image that was made available in October 2017, depicts an MRI (ww.pixabay.com)

now you can get an MRI, a CT, you can diagnose things that would be harder to diagnose before.”

“experience makes diagnosing diseases more accurate. Technology gives us more data to make a diagnosis with higher confidence,” wrote Yip.

These advancements have contributed to her job and contributed to the health of the patient, but even so,

As certain advancements in technology have become more

18 Computer Psychos


This image depicts a doctors stethoscope taken in 2016. (www.pexels.com)

time-consuming and repetitive, doctors are not able to see as many patients, and some of the data that doctors enter won’t be very helpful. Edison says, “We are using technology that is designed by engineers who don’t really understand

“Experience makes diagnosing diseases more accurate. Technology gives us more data to make a diagnosis with higher confidence” -Yip

As tasks take longer for doctors to accomplish digitally, the more doctors have to take home and complete. Edison says, “I have to come home and enter all my charts at home because I cannot do it all at work, but it never used to be that way.” This reliance on technology, the good and bad aspects of it, have also made it so that if something as simple as the Wi-Fi not working means that doctors cannot access their charts or see the schedule for the day, which makes their job harder. The advancements in technology from various fields of medicine are still in the process of being perfected. The technology so far is effective for now, but in the long run, technology will be revised and developed to balance time, patient and doctor satisfaction, and effectiveness. Technology will most likely never end in developments when it comes to medicine, but it will be refined over time. Until that time, doctors must adapt and use the tools they are provided and do their job to the absolute best of their ability.

what works best for us.” For the moment, not all technology is physician-friendly and helps doctors less than people.

19 Computer Psychos


As tasks take longer for doctors to accomplish digitally, the more doctors have to take home and complete. Edison says, “I have to come home and enter all my charts at home because I cannot do it all at work, but it never used to be that way.� This reliance on technology, the on how technological advanceThe first illustration depicts a pro/con comparison good and bad aspects of it, have also ments are affecting the medical field. The second illustration depicts a flow-chart made it so that if something as simple as showing the most basic and topical identifcation methods of somewhat beneficial the Wi-Fi not working means that doctors and mostly not beneficial cannot accesstechnology. their charts or see the schedule for the day, which makes their job harder.

The Pros And Cons of Technology

PROS

The advancements in technology from various fields of medicine are still in the

CONS

This image that was made available in October 2017, depicts

Due to the ease of sharing (ww.pixabay.com) medical records between process of being perfected. The technology so far is effective doctors, first impressions can for now, but in the long run, technology will bebe revised and made without meeting a developed to balance time, patient and doctor satisfaction, It allows for more data collecpatient and without professition, so more information is nal opinions. available to make more educated diagnosises. Doctor-patient relationships are weakened because of Advancements can make a screens, so it can be difficult to doctor’s job more accurate see when something is differafter adkustomg. ent. It is less likely to loose medical If a doctor struggles with records. typing, everything bcomes significantly slower. Imaging clarifyies the changing structure of internal organs. It helps make the diagnosis of a disease more accurate. Voice recognition allows doctors to transcribe in real time.

20 Computer Psychos

Using a screen during a consulation takes away from one-on-one relationships. Some programs are slower than actual handwriting. It takes time to learn how to use newer technologies.


“Experience makes Indentification Flow Chart diagnosing diseases more accurate. Is it a form of tehnology Technology gives us more data to make a diagnosis with Yes No higher confidence” -Yip

s an MRI

Is it time consuming?

Yes

Does not apply

No

Is the learning curve worth the outcome?

Does if have more benefits than the thing it is replacing?

Yes Yes

Beneficial

No

No

Meidocre

Probably Not Beneficial Sources: Dr. Virginia Yip, Dr. Flora Edison, Dr. Donna Ruesink

21 Computer Psychos


What websites does the user visit frequently?

yes

Entertainment

Fashion Display ads about fashion. What doews the user do?

Ignores the ad.

Do nothing.

22 Computer Psychos

Removes the ad.

Clicks the ad.

Remove the ad.

Show this ad again.

Display ads about entertainment. What does the user do?


Start Is the User Logged On?

t

er NO

Do not Record Data

Do not show ads like this again unless the user visits similar websites.

Display similar ads and see what the user does.

Google Ad Generator Thought Process This is a flow chart showing how a web browser knows which advertisements a user is most likely to click. This program learns over time to provide advertizements that appeal to you. This helps companies to sell a product by finding the right person to send the advertise to.

Sources “How Targeted Advertising Works.� The Washington Post, WP Company. Duhigg, Charles. THE POWER OF HABIT. Random House, 2014. 23 Computer Psychos


AI in Buisness and Marketing By Deven Sterzing When an advertisement appears, many people don’t think twice about it, but an enormous amount of thought goes into that advertisement. Years worth of data and click history compiled into an artificial intelligence (AI) program to present the user with an advertisement that the user is more likely to click. In business and marketing, AI is present in almost every company. Companies like Cognitive Scale create software for other companies to use, and such AI can be implemented to make somebody’s life at work easier, or it can go so far as to do their job and free them to pursue higher-level jobs. It’s located in all offices and greatly improves their life at work. AI is also used in marketing to give the user a targeted ad that the user is most likely to click on. “Because AI is machine learning and natural language processing, it [AI] starts with the data you give it,” says the CFO of Cognitive Scale Deborah Sterzing. An AI software is useful for near nothing when the company first implements it, but as the program gathers information, it becomes increasingly advanced. For example, if someone has never shopped at Target before, then Target doesn’t know anything about that person, but as that person visits more and more, it can send them targeted advertisements that can make them more inclined to buy from them. AI programs also learn from people’s actions online. Depending on what websites the user visits, the web 24 Computer Psychos

The above image is of a common workplace to discuss things in a buisness. flicker.com


browser can give the user better-targeted ads that they are more likely to click. As stated by Deborah Sterzing, “Marketing is about how you sell to your customers and what people think of you as your company is your brand.” If customers are shown targeted ads that they enjoy, they will have positive opinions toward the business and hopefully chose that business rather than the competition. “We have an app that we have produced for Macy’s, and that app is a game; it’s called Swipe n’ Shop,” Sterzing says. This app allows users to choose which dresses they like the most, so when they go to Macy’s, those dresses will be in the store along with other related dresses. If people know that they can go to Macy’s after playing this game, it attracts customers to purchase and increase Macy’s profit. This is how AI can help businesses gain money, but AI also helps within the company as well. “AI helps to serve your customers better, and the better you serve a customer the more loyal they are to you,” says Nomanive, the head of product development states at Cognitive Scale. The more the customer knows, the more they are able to make the right decision. John Richter, the head of product management at Cognitive Scale, says, “We try to augment the customer and the employees.” They can be augmented by being shown what they want to see and being given better choices. They are able to get cheaper, higher-quality goods. He also states that they use AI software to augment the employees. Aiding the employees with this software will allow them to work more productively and efficiently. For example, AI allows employees to make better judgments based on their actions in the past. The software also allows them to take in information faster than they would normally. It can show them a set of data about whatever the employee is analyzing that is most important to what he or she is doing at the time. This will help the employee to be more productive as well as more efficient. This then frees them up to do other things, along 25 Computer Psychos


with going home to rest so that they can be more productive the next time they are on the job. Cognitive Scale is an AI company that sells and creates AI software to sell to other companies. They are stationed in Austin Tx, and was founded in 2013. Their software can be used in a variety of ways, and the company sells to three main types of companies: healthcare, financial, and commerce companies.

They sell mostly to the insurance side of healthcare, and the Cognitive Scale’s software allows the insurance company to take in information about the most popular rates and cost of healthcare. This helps the company find the cost that will help them and that the customer is most likely to enjoy. In hospitals, this software is also used to keep track of the client’s medical records, helping the doctor to prescribe the right treatment for them. It serves the general insurance industry in similar ways, except it accumulates data about what kind of cars someone can purchase, how many accidents they have gotten into, the severity of those accidents, and the amount of money they can spend of insurance. The software will take in all of those factors and even more to give the insurance salesman the easiest and most effective way of selling their insurance to them. On the commerce side, AI helps to keep track of past purchases and patterns. The more patterns that emerge, the more the software can do for the company. For example, if a lot of people start buying yellow shirts, then the program will tell the company to start designing and creating more yellow shirts. The software will then see people’s reaction to the new shirts seeing what parts of the shirt work and what parts don’t, then telling the designer to make more of this kind of shirt. 26 Computer Psychos


“If a machine can learn how to do something, and it can do it faster, then it frees that person up to do something else,” states Nomanive. Many people think that AI takes away jobs, and the simple answer is that it does, but it also replaces the jobs. The person working at a computer all day typing out numbers will no longer have to do so. Instead they are free to pursue higher jobs and rankings in the company. “So if you take somebody who all they do is pay bills all day and one of the things they need to know every single day is how much they need to pay and by when do they need to pay it. So is a machine can figure out how to read the invoices, and they’re many kinds of invoices and it can find the amount and who to pay and it can find that out immediately and quickly, then it can make that person’s job easier.” Sterzing says. Looking ahead, AI will be able to read entire books and create accurate summaries. Already AI programs are able to read a company’s paragraph of feedback and implement the comments into its own program. The future of AI in business and marketing is truly exceptional, and it will only continue to grow. AI is able to help a company from the customer to the highest-ranking employee. It is able to produce highly specific, targeted advertisements that they are likely to click on. It can increase the production and efficiency of a company tenfold, and we have barely scratched the surface of the uses of AI in business and marketing. Who knows what lies ahead, but many people are eager to find out.

“The future of business and marketing is AI” 27 Computer Psychos


User Interface Inspiration: The Chatbot

The chatbot is a powerful version of AI that is becoming common-place in everyday life. Below you can learn about how a chatbot works and view the demo.

AT&T

10:10 AM ASK THE EXPERT

< Back

100% Messages

Hi, Welcome. What would you like to learn today?

By Tim Edison

Artificial Intelligence Machine Language Chatbots Chatbots OK. What would you like to learn about Chatbots? A History of Chatbots Chatbots in Our Everday Life Ode to Chatbots Chatbots in Our Everday life Chatbots play a vital role in many industries. Message based chatbots are growing in popularity by offering more convenience than visiting physical shops. This plays into the trend of the rise of online shopping, and combining chatbots with online shopping offers us... See More www.rilersguide.... Thank You so much! Have a nice day!

Ask The Expert left the coversation Text Messsage

Send

https://www.forbes.com

28 Computer Psychos


The Debate For and Against Chatbots What Is A Chatbot Pro (a) They provide ease of access. (b) Using a chatbot can free up time for a company. (c) Consumers don’t have to visit physical stores to get their items. (d) They reduce the use of humans doing menial tasks such as answering the phone. (d) Provides overall business efficiency

Cons a) They only have a limited vocabulary, so if it encounters something unfamiliar it will crash. (b) Consumer may need actual people to help them. The technology only goes so far as to provide basic information. (c) When a chatbot crashes, the consumer may have a bad time.

https://www.forbes.com

29 Computer Psychos


Siri, Live my Life By Tim Edison The alarm rings for the 20th time. Groggily pulling himself out of bed, Bob checks his phone and sees that it has prepared a list of all his meetings today. Upon entering his bathroom, Alexa reads him the day’s news on his command. After a quick shower, he hurries into the kitchen to make breakfast, accidentally spilling his glass of juice in the process. After mopping up the spill, Bob realizes that he is completely out of paper towels and calls out, “Alexa, place an order for a new roll of paper towels.” His Echo located on the dining table confirms his order and tells him that it will arrive later that day. Racing out the door, Bob remembers that he has a date that night and texts his local florist to place an order. An online chat bot greets him and places and confirms his order. Satisfied, Bob rushes off to work, knowing that today is going to be a good day. Bob is using Artificial Intelligence (AI) which is becoming

“Artificial intelligence to me is anything that makes our everyday life easier. Whether it be your phones navigator or something as simple as video games.”-Sam Jacob algorithms. Mr. James Shockey, the advanced computer science teacher at the Liberal Arts and Science Academy, describes an algorithm as, “a series of one, two, three steps.” Shockey describes a complex chatbot like Siri, an example of AI, as a “a vast number of algorithms

“Artificial intelligence is a form of intelligence formulated online that works through analysis and problems to increase its own knowledge. “-Joseph Carruth more and more prevalent in everyday life. Because it is a computer system mimicking human abilities, AI’s future is a gray space due to legal decisions. With computers growing faster in learning and storing previous information, debates about ethics and privacy concerns are igniting the technology industry constantly. AI swarms around an everyday consumer in three main forms-personal voice assistants, global positioning system’s

applied to answering questions.” Thus, AI is simply many algorithms working together to achieve a single goal which varies based on the nature of the program. With this basic knowledge of how AI works, it is possible to probe deeper into the above services.

(GPS) mapping user interference, and social media.

everyday consumer interact with daily. Joseph Carruth, a high school student, said that he uses Siri about once or twice a day, which is not as often as he should

The basis of all AI, from Bob’s Echo to his phone, is

30 Computer Psychos

Personalized voice assistants, such as Siri on Apple phones, are voice assistants (complex chatbots) that


device? After asking a few consumers of AI, the answer to both of these was unanimously--Yes. Carruth stated that he thinks that day is tomorrow; with constant innovation and development, he believes that, “we are going to allow our sensitive information to be accessed by an AI.” Jacob said similarly, stating that “almost all of these programs have allowed extensive tracking of information that cannot be erased, such as phone numbers, addresses, and bank account numbers,” which creates a big risk in personal privacy. It is likely that the ethics debate about AI will continue for many years, perhaps even until government intervention. The only thing that consumers can do for now is sit and wait. Another example of AI is the GPS device. GPS systems are built into almost every smartphone nowadays by major companies and are the basis for apps like

“An AI is a bunch of lines of computer code given meaning by the creator and then given knowledge by itself.” -Joseph Carruth

Sam Jacob, an avid user of Artificial Intelligence and an airman in the United States Airforce. (Photo Credit: Sam Jacob Photo)

Google and Apple Maps, which are

ing

With these advancements of AI, a growing

Al

Na

e

ug

a ng

l La

a tru

s hm

go rit

because he thinks it is such an intelligent tool: Siri has come a long way from unintelligent program to fully functioning question-and-answer system. The action behind how Siri, Google Assistant, and other speechanalyzing voice assistants is natural language processing. Its goal, as Shockey describes, is to find “how to parse through a sentence [and] how to extract the meaning of what we are talking.” Siri is ridiculed by some users for its nonsensical answers, but it has now developed a way to process the inputs that Bob gives it. For commands that it understands, it performs the corresponding function, and for nonsensical commands, they are defaulted to a web browser. Samuel Jacob, an airman in the U.S. Air Force, said that “the accuracy of Siri really comes down to what exactly I need. If it[‘]s something simple[,] it’s usually accurate. But if it is something complicated[,] she will just suggest Google.” Jacob feels like Siri is more of a convenience than a necessity.

s ces o r P

Code

concern among consumers is unfolding. Where will companies draw the line? Has the time come that information is going to be extracted out of one’s

31 Computer Psychos


The program Siri In action. Siri is a comlex chatbot that uses natrual language processing. (flickr.com/ Kārlis Dambrāns/ 2014)

really just GPS systems with user friendly interfaces. However, these days, map apps are becoming more than just GPS-based. Many are providing valuable information about roads, traffic, food venues, and what a destination is like. This is an incredible feat seeing as how they keep

maybe seven on recommending the best food places in town. Usually it just gives you whatever is closeby. So you have to google the best places. Which takes much more time and is inefficient especially if you’re driving.” Besides its downfalls, Jacob agrees that this information is extremely useful: “Smart navigation saves gas and time,

“Artificial intelligence...is really just trying to create algorithms in a computer that mimic how we make decisions” -James Shockey their data accurate. Google Maps isn’t perfect, however. According to Jacob, on a scale of one to ten, “it is at a six

32 Computer Psychos

and it could potentially help with traffic jams. I believe these resources are extremely useful.” Analysts project


Gooogle’s Personal Assistant, Google Home, makes a perfect addition to any room. The voice asistant is similar to Amazon’s Echo and Echo Dots (flickr.com/NDB/2017)

that Google Maps’s software to get better and better everyday, and with more people posting reviews online, companies like Apple and Google are obtaining vast amounts of data t. The future of mapping applications is, like most AI, is looking bright for many consumers. But probably most important, AI is revolutionizing social media and how we communicate. Many platforms such as Instagram, Facebook, and Snapchat can suggest followers or friends based off of contacts and viewed posts. Jacob said, “Honestly I think it is pretty brilliant. They suggest people who are friends with the people you interact with assuming you may or may not know them.” Carruth adds that he too thinks that the system is well-designed but feels “like it could be tailored a bit more by me and so as not to use an AI the whole time to

decide what I want to see.” There is currently no solution to this problem--it’s up to the producer of these products. AI is evermore playing a large role in our everyday life, and most experts at leading universities and Silicon Valley expect this trend to continue. Bob encounters it everyday in almost every app that he uses, and it is constantly improving the work that he does. The future of AI is now, and the decisions we make will determine the direction that the world of AI goes. And if Bob has a good date.

33 Computer Psychos


Video Game AI By Ikthan Marroquin Think of a video game so smart that it can challenge someone to their very limit while outsmarting them and learning from their movements, using them to defeat themselves. While this might sound like too much for a video game to do, there are some that have been doing this for years, with the help of artificial intelligence.

Users have different opinions about the current efficiency and realism of the AI in video games due to the multiple systems that have been used over the years. All these systems have new features and characteristics that the previous ones lacked. Among so many different systems, there are bound to be some errors.

Artificial intelligence (AI) has been used for the past few decades in video games to generate human-like

Luis Carlos Morillo, a high school student and a constant video game user, stated that AI, “can get to be pretty

Vintage controller. (Graphic/OpenClipart-Vectors) actions and responses in NPCs (Non-Player Characters) to make their behaviors become more user-friendly. All of this gives a much more immersive experience for gamers around the world. Companies know that to stand out they need to be different from the rest. This has led to the creation of different AI systems. These systems are usually different from each other, specializing in different aspects such as interactions or reasoning complexity. This allows the discovery of unique NPC personalities and features based on their persona and/or developing relationships with them.

High School student, Luis Carlos Morillo sitting in the Big Bend National Park, US, Texas. (AP Photo/Beatriz Rodriguez/2016) glitchy, like the typical moment in which you see a bot running towards a wall and when-face slamming against the wall still performing the running animation.� There are countless of such commentaries and reports of malfunction and illogical scenarios in video games in which artificial intelligence becomes more artificial than intelligent. Although developers work hard attempting to eradicate any possible glitch and/or bug, it seems as if every time one is fixed, another one pops out.

Mass Effect 3 Citadel AI in interaction with player. (Screenshot/Ryan Somma/2013) 34 Computer Psychos

Max Ramos, a high school student and a user of video games, gives a nearly perfect example of a scenario in which AI could really mess up. He says, “the point is that after saving [a city] from a DRAGON, [he] got arrested because a chicken decided to have a walk in the middle


of a fight between an [ice/fire] breathing dragon and a mighty warrior that could literally shout you apart.” As ridiculous as it may be, glitches and bugs in video games have been occurring for years during their existence. Although developers try to provide the perfect video game, this is much farther away than one might think. Errors aren’t something altogether very rare; people have really gotten used to witnessing a character accidentally morphing with another one or similar without thinking of it as something extraordinary. Ramos said, continuing,”So here we have two major logic malfunctions. First, the chicken did not react to the battle taking place causing it to simply carry on its normal walking animation. Second, I don’t really think one would get arrested for accidentally ending a chicken in the High school student middle of a battle against Ernesto Rios in Canada, a dragon in the process of which a city was saved.”

Quebec getting out of a store. (AP Photo/Di- Needless to say, artificial ana Trejo/2017) intelligence

may not always be as wondrous as one may expect, and while The Elder Scrolls: Skyrim is a highly acclaimed game having won multiple awards, it has some well-known, not-as-impressing AI responses. Errors in the behaviour

Dragonborn in battle, The Elder Scrolls V Skyrim. (BagoGames/2017) of NPCs, including their responses to dialogues and events, are nothing new to the gamer community. But just like there can be cons, there can be pros. If properly designed, AI can provide that great immersive experience so many users seek, having new systems developed that bring a new level to the abilities of AI. Sazi Marroquin, chief of edition at a news network said, “I believe that it [AI] has given

Shadow of Mordor, Talion in battle against orcs. (Andy Gilleand/2015) Assassins Creed 3, Connor running through a battle between Red Cloaks and The Patriots. (Tatiana T/2017). The Elder Scrolls V Skyrim, Book collectible along the game. 35 Computer Psychos


video games the chance to be more realistic [with] the different situations and the behaviour itself of the game taking algorithms that advance learning from themselves, and the decision-making of humans.” Video games such as F.E.A.R have developed an AI that is exceptionally good at adapting to the users’ surrounding and using it against them. It uses simple “taking-cover” mechanics to give the illusion of being flanked or being in the middle of some sort of coordinated tactic designed to ambush the character. If an enemy or group of enemies are taking damage and failing to take you down, they will grab their radios and call for reinforcements. Another great example of proper and innovative AI is the game Shadow of Mordor, which uses the AI system Nemesis System to give the NPCs the ability to remember previous events in which a user participated in. This system also generates a series of emotion-like actions coming from NPCs in the game.

Playstation 4 with controller. (BagoGames/2014)

But when do gamers know that they are in the presence of a truly good, smart AI? Assistant professor at the University of Texas, Alexander Huth, Ph.D., said that it “depends on the human standards at the time. Some decades ago, losing to the computer in a game of chess was quite impressive. Nowadays, that is just something that a computer does. Not a big deal.” There are a lot of new and exciting AI in upcoming titles such as Assassin’s Creed Origins, Far Cry 5 and Red Dead Redemption 2, all of which are sequels for what have been enormously successful franchises. They show promising, new and innovative gameplay and experiences that are sure to provide a step up in the video game industry and the AI in it.

36 Computer Psychos

Shadow of Mordor, Talion holding the head of an orc enemy. (BagoGames/2015)


Artificial Intelligence

37 Computer Psychos


Video Game AI Generates human-like reactions/responses in NPCs (Non Player Characters)

AI System Nemesis System

Shadow of Mordor

NPCs have memories of events and interactions they had, affecting their future desicions.

Radiant AI

The Elder Scrolls Skyrim

NPCs will react with emotions and change actions based on events taking place near them.

Assassin’s Creed III

NPCs react to current conditionw (weather/injuries), what they see, and other.

GDC Vault

The Perfect Organism Meta Stability 3.0

N/A

38 Computer Psychos

Characteristics

Alien Isolation

Halo Reach

F.E.A.R

NPCs can hera and smell. When player is hiding, NPCs will have an estimate of player location and then search. Enemies’ capable of evading incoming projectiles, besides reading your position and your health state. NPCs comunicate and cordinate different ways to engage by reading player position and fighting style.


Video Game AI Generates human-like responses/reactions in NPCs (Non Player Characters) providing a more inmersive experience.

Shadow of Mordor

The Elder Scrolls: Skyrim

Assassin’s Creed III

Alien Isolation

Nemesis System

RADIANT AI

GDC Vault

The Perfect Organism

Meta Stability 3.0

NPCs can hera and smell. When player is hiding, NPCs will have an estimate of player location and then search.

Enemies’ capable of evading incoming projectiles, besides reading your position and your health state.

NPCs have memories of events and interactions they had, affecting their future desicions.

NPCs will react with emotions and change actions based on events taking place near them.

NPCs react to current conditionw (weather/injuries), what they see, and other.

Halo Reach

F.E.A.R N/A NPCs comunicate and cordinate different ways to engage by reading player position and fighting style.

39 Computer Psychos


AI


Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.