Caleb Roberts Portfolio

Page 1

Portfolio Caleb Roberts M. Arch II Southern California Institute of Architecture



Portfolio Caleb Roberts M. Arch II Southern California Institute of Architecture



Contact

Education

Caleb Roberts

Southern California Institute of Architecture (SCI-Arc) . . . . . Aug ‘20 - Sep ‘21 Los Angeles, California Masters of Science in Synthetic Landscapes

Location . . . . . . . . . . . . . . . . . . Los Angeles, CA

Southern California Institute of Architecture (SCI-Arc) . . . . . Aug ‘18- Sep ‘20 Los Angeles, California Masters of Architecture II Clemson University. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aug ‘14- May ‘18 Clemson, SC Bachelor of Arts in Architecture

Email . . . . . . . . . . . . clbroberts24@gmail.com Phone . . . . . . . . . . . . . . . . . . . . +1 803-543-6477 Social Media . . . . . . . . . . . . @calebroberts24 Youtube . . . . . . . . . . . . . . . . . . . . Caleb Roberts

Southern California Insititute of Architecture Los Angeles, CA IDD & Seminar Teaching Assistant . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aug ‘19 - Dec ’19

Publications The Surreal Visions of Hernán Diaz Alonso - Jan. ‘19

CityX Venice - Venice Biennale - May ‘21 SCI-Arc LA Bestia - Apr ‘21 SCI-Arc Spring Show - May ‘19 SCI-Arc Obsessions in Focus - Apr ‘19

Expierience

Exhibitions

Teaching assistant responsible for giving software tutorials, conducting desk critiques, and collecting assignments.

Houdini

Cinema4D

Octane

Unreal Engine

Python

Machine Learning

ZBrush Maya

Portfolios

References upon request.

Stevens & Wilkinson Columbia, SC Intern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jun ‘17–Aug ‘18

Clemson SPIRO Clemson, SC Graphic Design Intern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jun ‘17–Aug ‘17 Graphic Designer responsible for creating branding material for student entrepreneurs and teaching students graphic presentation skills.

Awards & Recognition

VRay

Revit Photoshop Illustrator InDesign After Effects

Designed and visualized clothing, accessories, and furniture pieces.

Worked on DD and CD phases of three courthouses and 2 highschools, produced project renderings and VR experiences, Prepared 3d digital and physical models, developed and expanded the firm’s existing BIM library.

Proficiencies Rhino3D Grasshopper

HDA-X Los Angeles, CA Intern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . May ‘19 - Sep ‘19

SCI-Arc Continuing Scholarship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aug ‘19 SCI-Arc Admissions Scholarship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aug ‘18 ACME Composites Design Challenge (1st Place, Patent) . . . . . . . . . . . . . . . . . . . Apr ‘18 Clemson Merit Scholarship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ‘14 - ‘18



Table of Contents 3GAX

EDGE

DS 4000

Antrhopocenic Landscapes

09

DS 1300

Synthetic Landscapes 1

151

AS 2509

Details, Details

23

DS 5000

Synthetic Landscapes 2

159

AS 3222

Design Documentation

29

VS 2811

Rewilding Infrastructure

167

AS 2763

AI Assemblies

177

AS 2812

Ambiguous Forms

187

Arithmetic Imaging

195

2GBX DS 1201

Atlanta Public Library

43

VS 4201

Soundscapes

55

MArch.II Thesis

HT 2201

Cities of Celebration

63

DS 1420

DS 1200

BREAK AND ENTER

67

DS 1200

One-To-One

75

DS 1200

Visual Machine

89

HT 2200

A Stroll Through Los Angeles

97

VS 4200

Augmented Object

103

AS 3200

Tectonic Manipulation

113

DS 5000

Embedded Monastery

123

DS 5000

Convergent Horizons

131

DS 5000

The Perch

143

2GAX

3GBX


Caleb Roberts

Anthropocenic Landscapes

Umi No Mori Island in Tokyo Bay has been the primary landfill site for all of Tokyo for the past few decades. However, the island is now full and a decision is being made about how to use the land after the closure of the landfill. Many in Tokyo would like the island to be covered over with vegetation and turned into a pristine natural landscape, concealing—forgetting—the garbage beneath. This project argues that simply covering over the refuse will never be enough to turn this repulsive landscape into the natural landscape Tokyo desires. Utilizing similar strategies already implemented on the site and in other similar landfill reclamation projects—the introduction of foreign plant species, creation of a green landscape, and introduction of pedestrian circulation and community space—the project reverses many of the cities desires for the site: instead of a perfect, natural blanket of vegetation, the landscape becomes a gnarled and aggressive surface of invasive plants tangled with the garbage underneath. The two become one homogonous landscape, making it impossible to separate the natural from the artificial.

interchanges that appear on the map of the city—creating a link between these spaces and the infrastructural process and past of the site. These concrete pathways sometimes terminate in framed views of the landscape, bringing them uncomfortably close to the refuse that is the reality of the site. These three elements—the concrete spaces, the gnarled landscape, and the landforms made of garbage—create a reminder of the processes that created the island. Visitors are forced to see and experience the magnitude of the garbage present, created by them.

Situated within this rugged landscape, aggressive landforms bring the garbage out of the surface, creating an overgrown reminder of what lies beneath. Throughout the site, visitors encounter these landforms sometimes unexpectedly and are confronted face-to-face with the reality of the site and of the crisis that this garbage has created on a global scale. Visitors circulate through the landscape on a sunken level beneath the surface of the island within perfectly crafted concrete pathways and plazas. The forms of these concrete spaces is extracted from infrastructural ghosts of Tokyo city—silhouettes of major rail stations and highway

8

3GAX Vertical Studio

Instructor: David Ruy

3GAX


SCI-ARC

9


Caleb Roberts

Site Plan showing infrastructural silhouettes from map of Tokyo

10


SCI-ARC

Closeup textures from the site

11


Caleb Roberts

View under a bridge on the island

12


SCI-ARC

13


Caleb Roberts

View of concrete wall from within the landscape

14


SCI-ARC

15


Caleb Roberts

View from bridge entrance showing landform

16


SCI-ARC

17


Caleb Roberts

View from a ramp to the lower level of the circulation

18


SCI-ARC

19


Caleb Roberts

View from an outlook framing a landform

20


SCI-ARC

21


Caleb Roberts

Details, Details Team:

Kumaran Parthiban, Priyanka Rajani

At the scale of the detail, movement relies on the connection and interaction between individual elements to drive it. This project is an exploration of movement within the architectural detail. A built-in cabinet relies on the movement of individual rail-bearing connections to indicate and perform the movement. This cabinet consists of three main elements: the three drawers, the slide rails, and the cabinet shell. The rails are situated within the shell on angled planes that determine the precise amount of movement of each drawer. The drawers are then fixed to the rails along linear bearings. As the drawers slide along the rails, the angled shell pushes each drawer out to a different distance. The detail focuses on the interaction between individual elements that generate multiple movements. At the same time, the cabinet slides horizontally and the drawers open perpendicularly to that axis. This double movement allows the operator to simultaneously open all three drawers instead of needing to open each one individually.

22

AS Details Details

Instructor: Dwayne Oyler

3GAX


SCI-ARC

23


Caleb Roberts

ELEVATION 0’

1’

2’

PLAN 0’

1’

2’

Drawer Slide

ELEVATION 0’

1’

.75” Medium Density Fiberboard

2’

.25” Medium Density Fiberboard Box

Corian Inlay in Door Panel 2” CNC milled Mahogany Wood Door Panel

PLAN 0’

24

1’

2’


SCI-ARC

Cabinet in operation

25


Caleb Roberts

Cabinet in Operation

26


SCI-ARC

Drawer Slide .75” Medium Density Fiberboard Corian Inlay in Door Panel .25” Medium Density Fiberboard Box 2” CNC milled Mahogany Wood Door Panel

Steel section in Drawer Slide bolted to Drawers

Cabinet Sections

27


Caleb Roberts

Design Development: Atlanta Public Library Team:

Richard Mapes, Julia Pike, Kumaran Parthiban, Priyanka Rajani, Irvin Shaifa, Yash Mehta, Hiral Ahir

The design for the Atlanta Public Library by Richard Mapes and Julia Pike centers around floors generated from terrains and a Diagrid outer structural skin. This project continued the design through design development with particular focus on structure, environmental systems, and life safety. Structural details were developed based on real-world examples to define assemblies and construction methods within the design. The primary structure is a steel diagrid that supports the facade and floors. The floors are constructed from a rolled steel grid with insulation and hand-troweled concrete finishing. Additional finishes were chosen and developed for all surfaces within the building. Additionally, HVAC was planned along with passive cooling and lighting strategies. The central circulation ramp, which extends from the ground floor to the roof, acts as a solar chimney while vertical fins are located on all facades to control solar heat gain. Emergency systems such as a fire curtain for the central ramp, sprinkler systems, and fire-rated wall assemblies are integrated into the design to ensure code compliance and life safety. The design of all of these elements was developed in conjunction with one another and the overall deign was developed into a cohesive, detailed construction logic.

28

AS Design Documentation

Instructor: Herwig Baumgartner

3GAX


SCI-ARC

29


Caleb Roberts

Construction Sequence: Second Floor Structural Slab

Construction Sequence: Roof Finishes

30


SCI-ARC

Building Chunk

31


Caleb Roberts

Structural Details

32


SCI-ARC

The structural system of the Atlanta Public Library is comprised of three systems: The exterior steel exoskeleton, the typical steel girder floor structure, and the laser cut steel “egg crate” structure that supports fiberglass skylight elements. The steel exoskeleton acts as the primary structural element supporting the floor structure as well as the curtain wall facade. This skeleton contains diagonal bracing that folds along the pleated facade and provides lateral rigidity to the building. The structure of the floors consists of a ring beam and a grid of steel girders with tertiary light gauge steel framing and hand sculpted rigid insulation to negotiate between structure and finish floor. Tied within the floor system, skylight elements are supported by welded laser-cut steel “egg crate” structures that span the gaps in the floor and allow light to penetrate through the fiberglass lites. Transitioning between these systems are a series of welded and bolted joints that allow movement from thermal expansion or rigidity where needed.

33


Caleb Roberts

Integrated Systems Details: Lighting, HVAC, Life Safety

34


SCI-ARC

Because the building has a curtain wall shell, the only vertical elements large enough to contain systems are the circulation cores. Thus, the floor slabs contain a 3 foot cavity to accommodate all systems for each floor within the slab. Electrical, HVAC, Life Safety, Lighting, etc. all run through the floors resulting in a series of systems details that rely and react to the undulations and topography of the floors.

35


Caleb Roberts

Gallery Box Connection Detail

Gallery Box Connection Detail

36


SCI-ARC

Vertical Louvre Connection Detail

Roof Detail

37


Caleb Roberts

Passive Cooling Strategies

38


SCI-ARC

Daylight Analysis

Solar Heat Gain

39


Caleb Roberts

40


SCI-ARC

41


Atlanta Public Library

We have conceptualized the library of the future as a place for both experience and production. Currently, libraries contain vast amounts of information in various formats; however, those formats typically remain separate and force patrons to choose either or when deciding how to experience the information they are accessing. Our library removes the need to choose between media types by combining, overlapping, and hybridizing formats into new modes of experiencing information. This means that if a patron wants to learn about anatomy, they can request a paper book that has embedded links to online content that can be accessed through the library. They can also choose to view a holographic representation in one of the vr rooms or have a full scale 3d printed skeleton created in the makerspace.

The form of the library provides for varying levels of privacy within the reading spaces as the faceted walls create nooks and semi-secluded spaces for private study and experience. Three large surfaces of colored glass enclose the large reading spaces. Two of these look outward to the city while the third looks to a plaza created on the site. The contrast and transition from sheared, blunt surface to highly articulated geometry explores the increasing development of programmatic complexity in the Library typology. The library as a whole seeks to provide patrons with new experiences of information through the hybridization of media and the creation of spaces that promote exploration and creativity.

In order to make this experience possible, the library is divided programmatically into two types of space: reading room and production core. The reading rooms are large multi height spaces set behind colored glass that provides an ambient atmosphere that changes throughout the day. The production core is a mostly enclosed space that houses the library’s books as well as all of the technologies that create the hybrid medias. Large volumes inside the library create production spaces for full scale robotic fabrication machines as well as digital media servers. The two types of space are separated by a thick shell; however, along the circulation lies an interface that allows patrons to see the encoding and fabrication of various media types as they move through the building. Patrons interact with these interfaces and receive information about what is being produced in the core as well as how they can experience more of what they are looking for.

42

2GBX Design Studio

Instructor: Herwig Baumgartner

2GBX


43


Caleb Roberts

Interior view of main reading room

44


2GAX

45


Caleb Roberts

North Elevation

South Elevation 46

West Elevation

East Elevation


2GBX

5th Floor Plan 10th Floor plan 1/16” = 1’

Ground Floor Plan Ground Floor Plan Ground Floor plan 1/16” = 1’

47


Caleb Roberts

Axon

48


2GBX

View from Auditorium Mezzanine

View from Plaza

49


Caleb Roberts

Model Photos

50


2GBX

51


Caleb Roberts

Interior view of escalators to main reading room

52


2GAX

53


Soundscape

This seminar focuses on a problem of formally representing knowledge in the age of the “big flat now”. If the classical ideas of ordering, indexing and systematic and functional organization were tied to the idea of encyclopedia as the possibility of fully describing, understanding and ordering the world, the seminar shifts its focus towards ideas of accumulation and aggregation as a new formal language that conceptualizes the internet as a new model of knowledge. This project utilizes a range of digital workflows (from sourcing of 3D models, to texture mapping to coding in C# and real-time rendering) utilizing a range of softwares (Internet object collections, Unity-Platform Sandbox, Maya). A focus is placed on sourcing

54

objects and designing a personal, custom made software imaging application in order to establish and automate a range of aesthetic microgenres within a common search space of a tool. The seminar also explores simulation as an aesthetic problem and treats screenshots as finished images in order to propose novel paths for digital design procedures as cognitive labor. A massive catalog of digital models is produced in order to establish a collective visual ‘search space’ and to think what it means to evaluate and select on such a large scale.


55


Caleb Roberts

The focus of this iteration is the music object. Each object’s role in and relationship with the world of music is intensely unique; however, when they are multiplied and stacked together, the significance of each individual object is removed creating a new, purely formal expression of the objects.

Melody

56


2GBX

Grunge Music

57


Caleb Roberts

Distortion

58


2GBX

Each composition recreates a musical concept in a purely visual manner. Taking advantage of impromptu pairings and intersections of objects by Platform Sandbox, these compositions formalize musical language into a visual expression of the concepts. Objects, piles and scenes are generated and create new overlaps of object and meaning.

Cacophony

59


Caleb Roberts

60


SCI-ARC

61


Cities of Celebration Ritual as an Informal Sapacemaker

On August 25th, 2018, nearly seventy thousand descended upon the desert of northwest Nevada into a temporary city larger than many well established communities across the country.1 At the same time, nearly five hundred thousand congregated, as they do every weekend from August to December, in the sprawls of parking lots surrounding football stadiums of seven major cities across the nation in 2 support of their favorite NFL teams. In the following month, the small town of Clemson, South Carolina increased its population by one hundred thousand in anticipation of the first home game of the season. Tents and improvised shelters popped up in every vacant space on the university’s campus while fans gathered in communal celebration of the game day events. Rahul Mehrotra in his essay, “Negotiating the Static and Kinetic Cities” explores similar circumstances in Mumbai, India during the Ganesh Festival where he sees a distinct divide between the Static architectural backdrop of the city and the Kinetic cultural inhabitation of urban 3 spaces. These two events, Burning Man and the American Football tailgate, exemplify a contrasting capacity for informal space making through the ritual of celebration that necessarily incorporate both the Static and Kinetic into one culturally cohesive space. The two events generate spaces that resemble small cities and communities

in that they contain places of commerce, temporary shelters, and community spaces. In these temporary informal urban environments, culture and economy are one and the same. The static environment becomes the backdrop for the temporary conglomerations of tents, trailers, and lawn chairs while the formal economy is subsumed by the cheaper, more convenient informal market stalls and street vendors. These informal environments share many similarities with the informal cities that arise in Mumbai during the Ganesh Festival and other cultural celebrations. The cities appear overnight and become sites for rapid cultural exchange then disappear again just as quickly as they are created. As Rahul Mehrotra explains, “there are no static or permanent 4 mechanisms to encode this.” The kinetic and ephemeral nature of these environments makes tying them to any constant within the static city quite difficult. These cities are better understood as an “enacted process” without any link to the physical constant of the static city. In fact, the only consistent measure or memory of these cities lies within social media where an accumulation of thousands of images and posts leaves behind the history of the events. In a publication for the Journal of Consumer Research, anthropologists Bradford and

1. Adams, Chris. Black Rock City Census. (January 2019): 12. 2. 2018 NFL Annual Attendance Survey. ESPN. Accessed March 12, 2019, http://www.espn.com/nfl/attendance. 3. Rahul Mehotra, “Negotiating the Static and Kinetic Cities”, in Other Cities, Other Worlds,” Andreas Huyssen (Durham and London: Duke University Press, 2008), 205-216 4. Mehrotra, 208

62

HT 2201

Instructor: Erik Ghenoiu

2GBX


Sherry offer an alternative term for the celebratory placemaking of these events.5 Vestaval, a term they derived from the words festival and Vesta, the Roman goddess of the hearth, foregrounds the communal nature of these informal environments. In the case of the tailgate, the informal environment takes the form of a neighborhood where each individual tailgate acts as a temporary home with spaces set apart for the primary functions of the home: kitchens take the form of grills and barbeques, living rooms become clusters of lawn chairs huddled together, and the patio or lawn can even be seen in the play spaces between the tailgates. Neighboring tailgaters often interact, sharing food and bonding through mutual team allegiance. Thus far, the three informal environments of the tailgate, the festival in Mumbai, and Burning man have been equated and their similarities generalized. However, a particular distinction must be made about the Ganesh Festival and the Tailgate. The American tailgate assumes an existing urban condition to provide facilities, infrastructure, and a built environment--after all, the culminating event is a football game taking place within a mega structure. While the Ganesh Festival takes place in an urban environment, it does not necessarily rely on or interact with the Static setting. The primary event of the Ganesh

festival is the culminating procession of the idols carried to the sea. Leading up to this event, communities create temporary public spaces to house the idols with the static city as a spatial delineation.6 The procession then moves through the static city, but does not require or necessitate interaction with it. In the case of the tailgate, the informal communities rely on the static organization of the university or city to determine their boundaries. Lawns in front of lecture halls, plazas outside of academic buildings all become filled with these temporary neighborhoods.7 Sometimes, the static architecture also plays a part in the informal events: university buildings are decorated in school colors with banners and lights, facilities are opened to the public for the duration of the game. In most contexts, as described by Mehrotra, the informal ritual city does not interact with the formal static city. The separation is, to some extent, a result of impermanence of the ritual celebration. The festival or celebration does not last; and when it is over, the informal city dissipates allowing the static city to regain its dominance. However, an interesting hybrid spatial typology can be created where the two environments collide. This typology employs both the informal and the formal to initiate and promote interaction

5. Tonya Williams Bradford, John F. Sherry, “Demonstrating Public Space through Ritual: Tailgating as Vestaval.” Journal of Consumer Research 42, no. 1 (2015): 130-51 6. Mehrotra; 208 7. Bradford; 135

63


Caleb Roberts

between the Static and Kinetic. The Los Angeles Rams stadium designed by HKS is an example of this collision. It is designed in a way that forces the static architecture to play a role in the informal experience of tailgating. Spaces under its large roof are allocated to the tailgating experience allowing fans without a ticket to the game to participate in tailgating activities within the stadium. The space is designed to be used in the same way tailgaters use empty lots promoting interaction between the formal and informal environments. Spaces that interface between the informal and the formal, such as that in the Rams stadium, can be placed into a category in between the informal and the formal that integrates the two. Hawkins and Ryan, Applied Business professors at the University of Tasmania, call these spaces “third places” where the first two are the static and kinetic. Hawkins and Ryan identified distinct aspects of each type of space in The Falls Music and Arts Festival in Tasmania, Australia.8 The festival takes place in a semi-permanent development constructed specifically for the festival. A mixture of permanent infrastructural elements and temporary communal concert, art, and dining spaces so that the formal and informal imbricate into a third place. They identified these third places as sites for

interrelationships between informal leisure, social capital and place characteristics and they found that the success of these places was determinate of the success of the festival. These examples are just a few of the hundreds of informal environments generated by ritual. Each manifests differently within the given cultural context and each relates uniquely to the static city. Despite the various levels and modes of integration between the informal and the formal, each of these cases displays the potential of the ritual to create complex interweavings between the informal and formal environments.

8. Clayton Jon Hawkins, Lee-Anne J. Ryan, “Festival spaces as third places”, Journal of Place Management and Development, 6, no. 3 (2013):.192

64

Essay Topic

Informal Cities

2GBX


2GAX

Top - Ganesh Festival, Mumbai;

Source: theindianawaaz.com

Middle - Burning Man, Nevada;

Source: journal.burningman.org

Bottom - Clemson University, South Carolina;

Source: fansided.com

65


Break & Enter Partner:

Kshitij Bhende

We believe that all architecture should belong to the public and that no entity should be allowed to mediate public access. However, there are currently dozens of buildings in the LA area being held captive by needless regulations such as open hours and entry fees. We believe these buildings should be returned to the public; and as such, we have performed a series of carving exercises that provide a new means of access to these architectural marvels. We are currently working on numerous buildings in the LA area. However, we will focus today on the Bradbury building. Our intervention is a simple carved hole and inserted fiberglass element. It is expressionless in materiality, but active in form, relating to the larger context of our work. It brings into the foreground the greater vision of public access to all architecture. Through this act, we invite the public to once again enjoy and benefit from the architecture that for too long has been locked away from them.

66

2GAX Design Studio

Instructor: Natasha Sandmeier

2GAX


Image of an Intervention at the Bradbury

67


Caleb Roberts

Dozens of works of Architecture throughout Los Angeles limit public access unnecessarily robbing the community of valuable cultural and educational experience. These interventions circumvent these institutional regulatory acts, giving the buildings back to the public. Each intervention is designed and fabricated specifically for its intended building

Chinese Theatre

Broad Museum

The Getty

LAX Theme Building

A collection of interventions throughout LA

68


2GAX

Cal Trans

Capitol City Records

Chateau Marmont

Ace Hotel

Walt Disney Concert Hall

Griffith Observatory

Stahl House

LACMA

69


Caleb Roberts

Shop Drawings for The Bradbury Intervention

70


2GAX

71


Caleb Roberts

A Photo of the fiberglass forms in the warehouse

72


2GAX

73


One-To-One Partner:

Kshitij Bhende

One-To-One explores the dissociation of architectural form from its context and the potential that creates for new relationships with new programs and adjacencies. Service and public circulation elements of the Bonaventure Hotel were extracted through 3D scanning technology and the resulting forms recombined and merged while maintaining scale into a new form for use as a flower shop on the Venice Boardwalk. By realigning the one-to-one relationship of these parts and embracing the imperfections of technique, the architecture is allowed to create a new language and expression while, at the same time, maintaining an element of familiarity to its origin.

74

2GAX Design Studio

Instructor: Florencia Pita

2GAX


Resin Cast detail of the flower shop

75


Caleb Roberts

Collage of a service stair somewhere at the Bonaventure

76


2GAX

Collage of a mechanical unit somewhere at the Bonaventure

77


Caleb Roberts

Collage of the Tower 3 stair in the Bonaventure

78


2GAX

Collage of another service stair somewhere at the Bonaventure

79


Caleb Roberts

Flowershop Model with all textures and inaccuracies from scanning

80


2GAX

81


Caleb Roberts

Flower Shop Elevation

82


2GAX

Flower Shop Elevation Detail

83


Caleb Roberts

Flower Shop Plan

84


2GAX

Flower Shop Plan Detail

85


Caleb Roberts

The Flower Shop exists as a result of an overlap of digital and physical production methods. Digitally scanned and composed details of the Bonaventure are 3D printed and dipped in resins by hand to produce variation of surface articulation from rough to smooth, matte to glossy.

Flower Shop Model

86


2GAX

The Flower Shop creates formal relationships with its context that allow for new circulation paths and points of rest along the Venice Board Walk. It is similar in scale and function to the tents of street vendors while adding a contrasting aesthetic to the grunge and informal space of the street market.

Flower Shop Model

87


Visual Machine Partner:

Kshitij Bhende

In his 1996 publication entitled Interrupted Projections, Neil Denari discusses the encoding and oversaturation of visual language and sign-forms on society. During the time he was writing, the predominant visual language was, to him, the logo. Today, the ubiquity of social media has shifted visual language to the image. The Visual Machine is a mechanical apparatus which displays and exaggerates the encoding of the image; the end result is an over-saturated and sometimes unrecognizable display of the image encoding process. New landscapes are generated by the data that forms an image. The machine is experienced from a separated platform giving individuals the same detached and impersonal experience of the displayed image as they would receive from social media.

88

2GAX Design Studio

Instructor: Angelica Lorenzi

2GAX


89


Caleb Roberts

Section Image of the Visual Machine

90


2GAX

91


Caleb Roberts

Images fed into the machine are reduced into the most basic form: pixels rgb. Mathematical representation of color value triggers height variations that disintegrate the traditional method of experiencing imagery and immerse the viewer into an unfamiliar landscape formed from an image of a place already known

Closeup of the Visual Machine in Operation

92


2GAX

Images of the Visual Machine Gallery experienced in AR

93


Caleb Roberts

Closeup of the Visual Machine in operation

94


2GAX

95


A Stroll Through Los Angeles An analysis of one possible future of the city

As the train approached the station, Alex heard a woman’s voice over the intercom, “Next Stop: Union Station, Los Angeles,”. He stepped lightly off the train onto a bright, crowded platform. He had read an article recently in which he learned that, some years back, LA city developers had abandoned housing development for more lucrative data farming. As he walked through the congested space of the station, he noticed that nobody was leaving the station. It seemed that every traveler was only transferring to another train. Union station was no longer a destination, but more of a midpoint that collected passengers from hundreds of cities and redirected them toward the next leg of their journey. After searching for several minutes, Alex found an exit sign pointing toward a dimly lit passage. As he walked down the hall, Alex heard a voice behind him, ”Are you lost? Nobody ever leaves the station anymore.” Alex turned around to see a very rugged man dressed in grease-stained coveralls carrying a toolbox in one hand and a copy of the morning’s newspaper in the other. The headline read, “E-Residency: The Next Craze in the Internet Boom”. Alex thought about the question briefly then answered, “No. I think I am exactly where I need to be. I’m a PhD student studying how all of that,” he gestured toward the newspaper, “is transforming our society.

96

HT 2200

What is the future of the city? One possible future sees the city cannibalized by infrastructure in order to meet rising needs of society. The city of Los Angeles reacts to production, transportation, and data needs of the near future and becomes a post-human framework for the world’s economy. In the current state of the city, this future seems an extreme leap. To understand how this future might come to be, we must look to the past and the present to uncover the developmental trends and importance of infrastructure to the city and to society. Alex’s first encounter with Los Angeles Infrastructure is in Union Station where he realizes that nobody is leaving the station to go into the city. For the traveler, the city becomes merely a weigh-station, a transition point between one leg of a journey and another. In this future, the human has no place in the city outside of this rapid, fleeting interaction. All other existing transportation infrastructure is re-purposed into another step in the automated assembly line of production. The streets, highways,

Instructors: Marcelyn Gow & Tim Ivison

2GAX


To get the best understanding, I figured I should start at the source, Los Angeles.” He turned to walk away, sensing that his answer sufficiently satisfied the man’s curiosity. However, after only two steps, the man’s voice started again, “Well, you’re going to need someone to show you around the city then, aren’t you? They stopped updating the Google Maps of LA when the last residences were converted and I’ve been a maintenance worker here since Google took over the Fashion District.” Alex waved his hand, gesturing for the man to join him, “My name is Alex, by the way”. After tucking the newspaper under his arm, the man reached a gnarled, calloused hand out toward Alex, “Call me Steve,” he said as they walked toward the entrance of the station. The automatic doors grinded open—it was obvious that they did not get much use anymore—and Alex was immediately immersed into a foreign world so alien to him that it stopped him in his tracks. Automated vehicles zipped by with great speed and efficiency, ‘drones flocked the skies carrying Amazon packages, the city set behind appeared more like a jungle overgrown with fiber-optic wires, conduit, and radio antennas. “Ten times as many vehicles on the roads now than when people lived here and yet I haven’t seen a single wreck since I started working,” Steve broke the silence. “How do they not run into each

SCI-ARC

and railways of the city become the arteries through which pulse with the necessary goods to keep society alive. Early proposals for the interstate and continental rail systems portrayed these infrastructural elements in this way: efficient, machinelike appendages that connect and streamline economic actions to the individual and to society. Keller Easterling explains that, in the early-to-mid 20th Century, the development of transportation infrastructure was predominantly focused on the configuration and recombination of intermodal transportation networks that overlaid and intertwined human transportation 1 with that of production. In the new city, however, a distinct separation prevails as a predominant organizer of space and settlement as production becomes the focus and automation allows society to live independently of those processes which allow it to function. Beyond transportation networks, data and information become predominant spatial organizers of the new city, entangling and enfolding the urban landscape into a wholly

1. Keller Easterling, Prospecta 30: Interchange and Container, (Connecticut, The Yale Architectural Journal, 2003), 113-115.

Location

Los Angeles circa 2075

97


Caleb Roberts

other?” inquired Alex. “They ripped up all of that wretched pavement at the beginning and installed thousands of sensors in the roads. The cars all communicate with each other now.” Steve pulled a phone out of his pocket and tapped a few times on the screen. Immediately, one of the cars came to a quick stop right in front of us and the doors opened. Alex noticed that this car looked different than the others on the road. It was smaller and less boxy. Upon his inquiry, Steve told him that this was one of the few vehicles still meant for human travel. They were made for the maintenance workers who periodically come to the city to monitor system performance. All of the others are autonomous freight vehicles owned by the various large corporations. The only thing they carry now is merchandise. The two stepped into the car and it zipped back into the current of moving vehicles. The two had agreed that Alex should visit One Wilshire, the place where the city’s metamorphosis began. Along the way, Alex stared out the window at the alien world passing him by. He could almost recognize some structures as apartments, office buildings, shopping malls. Still, much of the city’s past was already masked and consumed by layers of cables, cooling pipes, transistors, relays, and antennas. In some places, the labyrinthine mesh of

2.Kazys Varnelis, Blue Monday: Ether One Wilshire, (Barcelona, Actar Editorial, 2007), 48-62.

98

connected system—not just internally connected, the system spans as far as the satellites orbiting in space can throw their web of radio waves. Common infrastructural systems are re-purposed into new networks that better serve the infrastructural city. Fresh water lines are replaced with cooling systems and hydraulic networks to propel and maintain the vast amounts of systems and functions of the mechanical and informational structures. One Wilshire is at the center of this fabric in Los Angeles as it was the first in a long line to displace human habitation in favor of a more lucrative occupancy of information. Along with describing a new world order in which communications technology dominates, Kazys Varnelis points out various socio- and geopolitical processes which led to One Wilshire’s status at the forefront of this evolution: The cold war, the space race, the rapid privatization of the internet all situated One Wilshire as a powerful and profitable site for 2 infrastructure’s sublimation of the city. Not only was it more profitable, but it was also more efficient than clogging


the infrastructural network enshrouded the existing buildings, resulting in a purely mechanical scene that looked like something from The Terminator. In other places, the network was laid more intricately into a sort of filigree which traced the details of a few Art Deco and Post-Modern structures, like the US Bank Building, creating a delicate pastiche of the old and the new Los Angeles. Alex found it quite unsettling that the cultural melting pot of LA that he had read so much about could just be covered up by this cold, unfeeling, and efficient machine of a city. “You get used to it,” Steve broke the silence. “It helps to not think of Los Angeles as a city anymore. It’s more of a giant server farm now”. Alex thought about that for a moment and responded, “So the entire city is like this? Servers and satellite dishes everywhere?”. Steve nodded, “Pretty much. Near the port, it turns into warehouse storage and shipping yards. They put in automated rail stations and dry-docks for loading and unloading goods to be shipped out to other cities around the country. Los Angeles used to be home to Hollywood and the Rams; now its home to the internet and the largest port in the world.” As the journey continued, Alex imagined that all of the cables spilling from every apartment building and residence traced the daily path of someone who lived there before. Before he knew it, the car jolted to a stop

SCI-ARC

the system with human interaction an error; thus, the entire city followed suit, giving way to an urbanity totally enthralled by infrastructure. As Alex moves through this unusual urban landscape, his encounters with the technology of the new city disturb and disconcert his understanding of the relationship he, and society as a whole, has with technology. Adam Greenfield and Liam Young both chronicle a future of technology that is not only completely alien, but 3,4 also wholly intertwined with society. Surely this connection between human and machine will persist into the new decentralized version of human settlement that will arise in this future, but it is beyond the scope of this essay to explore outside the bounds of the city. But what if technology becomes independent? Greenfield and Young both speak of the emergence of machine learning and intelligent systems. Yet, in their futures, technology still operates very traditionally, still requiring a human operator. Within the new city, the development of increasingly efficient automated systems nullifies the

3. Adam Greenfield, Radical Technologies: the design of everyday life, (London, Verso, 2018), Introduction 4. Liam Young, Peter Frankfurt, and Sou Fujimoto. “The Other Future.” Lecture, Designing the Future Symposium, Sci-Arc, Los Angeles, October 25, 2018

99


Caleb Roberts

and Alex could see what he assumed to be the One Wilshire tower out the window. Stepping out of the car, Alex’s hair stood on end and he could feel the air hum as the super computers within the building efficiently crunched the data and relayed the info of the entire world through wires and processors before sending it all the way through the fibers or satellite dishes out of the city and off to its final destination in the next city or thousands of miles across the world. “So this started it all, huh,” inquired Alex. “Yes sir, One Wilshire was the first telecom hotel, the most expensive property in Los Angeles. Now, it’s the heart of this whole operation. All of the information that passes through the millions of servers in the city, all of the commands that drive the machines in the ports and warehouses, everything is monitored and distributed from right here. I guess you could say One Wilshire is the command center of the whole city”. In the car on the way back to Union Station, Alex pondered what he had experienced. As perplexed as he was, he came to the conclusion that at some point it became more efficient for companies to have one centralized location for the infrastructural elements of their businesses and that, because it was already the headquarters for numerous major corporations and because of its situation as a major international port/ weigh station, Los Angeles was the most suitable location for such a city. At the station again, Alex bid Steve farewell and thanked him for accompanying Alex through the city. He made his way through the bustle of the station to the train heading toward San Francisco and returned home to begin his writing.

100

requirement for the human in the regular operation of the system. We still see that some human interaction may be required to maintain and repair the system. Whatever the future may be, it is evident that infrastructure and technology will undoubtedly spur transformations of society into something entirely new. Dependence on technology inevitably will push society forward. Whether that be toward an interconnected, involved urban environment where community and culture retain a symbiotic relationship with infrastructure, or toward a future where automation allows society to dissociate itself from the processes and frameworks that support and maintain it, cannot yet be determined.


101


Augmented Object Partner:

Phillip Hood

The Augmented Object is an exercise in manipulation and representation of a known object with 3D modeling and AR visualization to generate new relationships between nature and technology. 3D scans of a flower were morphed and voxelated then sliced and overlayed with images of the real flower in augmented reality apps that collapsed the two modes of representation into one plane. The intertwining of nature and technology, digital and physical separates the original flower from its representation to an extreme without it becoming fully unrecognizable.

102

2GAX Visual Studies

Instructors: Casey Rehm, Andrea Cadioli

2GAX


3D Printed Flower


Caleb Roberts

3D scan of a flower

104


2GAX

Anaglyph of the voxel flower

105


Caleb Roberts

Flower Aggregation

106


2GAX

107


Caleb Roberts

AR augmentation of flower

108


2GAX

Different modes of representation are collapsed into one interactive image that allows the flower to be read in new ways and understood differently than before.

AR augmentation of flower

109


Caleb Roberts

Anaglyph Overlay generated in After Effects

110


2GAX

111


Tectonics Feltrinelli Porta Volta -Herzog & De Meuron Team:

Ovgu Nurozler, Phillip Hood, Kshitij Bhende

This project was conducted in two parts. First, was a close analysis of Herzog & De Meuron’s Feltrinelli Porta Volta in regards to its tectonic expression and assembly. Particular attention was given to the 48° angle skew to the building’s structural grid and how this affected the structure and enclosure geometrically and tectonically. The Second part of this exercise focused on re-imagining the concrete building as a timber frame structure in an effort to understand the consequences a different structural system would have on the building’s tectonics and aesthetics. The primary focus was placed on joints and connections as well as maintaining as much of the original geometry as possible.

112

Tectonics

Instructors: Maxi Spina, Randy Jefferson

2GAX


113


Caleb Roberts

Corner Chunk

114


2GAX

The pinnacle of the roof presents a unique situation for the enclosure of the building. Because the glazing is carried to the top, a custom panel was fabricated to create a cap and seal the connection between the two pitches. On top of this, the precast concrete cladding rests to continue the angle to its point.

Typical Roof Detail

115


Caleb Roberts

Building Chunk after Timber Conversion

116


2GAX

At the level where the pitched roof begins, the structural elements are angled on two axes which generates a unique connection for the structure. The floor slab continues to cantilever outward from the facade while the glazing units become skewed parallelograms. The finish of the glazing also changes taking into account a more direct angle with the sun. Floor-facade detail at Pitch

117


Caleb Roberts

A collection of Timber-Steel connections

118


2GAX

These connections explore the necessary changes in geometry that must be made in order to accommodate for a timber structural system. Each connection joins a series of timber members with a welded steel bracket designed for a specific location within the structure. The use of steel connections allows for stronger joints that maintain as much of the original geometry as possible.

119


Caleb Roberts

Model of timber structure conversion

120


2GAX

Model of timber structure conversion

121


Embedded Monastery Arkxsite Monastery Competition

The aim of this project is to create the least visual impact on the natural environment and the historical ruins. A small dirt path meanders along the cliffside through a serene, powerful natural scenery to a place where the ruins of an old fortress lie. The fierce landscape and stoic ruins beg not to be disturbed. The monastery is embedded into the cliff, concealing itself within the rock, surrendering to the status and memory of the Baralha Fortress. Walking toward the site, one might not notice the entrance to the monastery tucked in the hillside along the path; and from the ocean, only a small incision into the landscape can be seen.

the existing path, or to descend into the landscape and into the secluded contemplation spaces of the monastery. The stair is situated just inside the slit in the cliffside offering views of the ocean along its length. Each of the programmed spaces is situated along the stair creating a simple sequence of movement throughout. All aspects of the design work to maintain the site’s existing serenity with the greatest respect for the memory of the ruins and the supremacy of the pervading nature.

The monastery takes care to make itself hidden while offering visitors of the monastery a quiet, still environment for reflection and meditation. The program is organized linearly along the stair that connects from the top entrance to the lower observation point facing the ruins. Visitors of the ruins may choose to walk along

122

3GBX Vertical Studio

Instructor: John Enright

3GBX


123


Caleb Roberts

View Looking Down Stair

124


3GBX

Elevation View from Ocean

125


Caleb Roberts

126

6. Meditation Space

3. Dining/Meeting Room

7. Lookout

4. Storage

8. Ruins

2.

1.

1.

1.

1.

2.

5.

2. Bathrooms

3.

5. Kitchen

4.

1. Dormitories

1.

4.

View from Dining Space


7.

6.

8.

3GBX

127


Caleb Roberts

Longitudinal Section

Transverse Section 128


3GBX

Roof Plan

129


Convergent Horizons Urban Confluence Silicon Valley Competition

The challenge of creating an iconic form for Silicon Valley within the low height limit of 200 feet is addressed in this proposal by a figural from that allows for multiple viewing points. Typical observation towers, the Eiffel Tower, Space Needle, and even the Wheel in London soar to heights above 400’ and can reach up to 1000’+. Therefore, to create an iconic form that cannot be taller than even the tallest building in San Jose requires a rethinking of the typology of the observation tower.

The tower is clad in a perforated metal skin with a programmable LED mesh concealed within, turning the skin of the tower into a performative element with patterns, messages, and designs shining through the skin. At night, the skin glows with moving light that does not negatively affect local and migratory wildlife. This skin leaves the possibility open for future community engagement in designing what is displayed with the lights.

The tower combines rethinks the traditional Due to the imposed height restrictions, the form concept of monumental observation tower of this observation tower breaks the traditional and combines this new, convergent viewing typology of observation towers that loom high apparatus with dynamic lighting and community above the city with a singular viewing platform. space to give both Silicon Valley and San Jose Instead, this tower takes a crystalline form and a truly iconic landmark. responds to the height limits by branching off into several arms that each frame the context uniquely. Each arm points toward a different horizon of the context creating four dynamic views of earth and sky that will undoubtedly change and develop as the city grows. These arms converge on a central viewing space within the structure that provides visitors views out through the arms as well as a space for gathering within the structure itself.

130

3GBX Vertical Studio

John Enright

3GBX


131


Caleb Roberts

View From W Santa Clara Street

132


2GAX

133


Caleb Roberts

Structural Axon

134


3GBX

Exterior Perforated Metal Panel

LED Mesh

Secondary Structure

Primary Structure

Interior perforated Metal Panel

135


Caleb Roberts

200’

160’

95’

40’

Section 200’

160’

95’

40’

Elevation 136


3GBX

View From Observation Deck

137


Caleb Roberts

Observation Deck Floor Plan

Ground Floor Plan 138

Ground Floor Plan


3GBX

View From Arena Green

139


Caleb Roberts

View from Inner Observation Platform

140


2GAX

141


Caleb Roberts

The Perch

Monte D’Oiro Wine Tasting Room Competition

The wine tasting room sits, perched at the edge of the winery, leaning out over the hill toward the vineyards, marking the corner of the cluster of existing buildings. Its four large apertures stretch outward and upwards directing visitors’ gaze to the rows of grape vines while maximizing the natural light in the space. The minimal, white interior serves as a backdrop for the experience of the wine and the views towards its origins. The tasting room is elevated by three legs which connect the upper areas of the site with the lower area of the vineyard. The building thereby acts as a kind of portal to the vineyard. The exterior of the tasting room is clad in corten steel, referencing farming machinery that occupy the rural area of the countryside. In contrast, the interior is composed of white marble and corian creating a light, airy space that fades to the background as visitors drink their wine and gaze out to the vineyards. The minimal interior finishes create a contrast to the intensity of the beautiful landscape, framing and enhancing the views of visitors as they enjoy casual conversation, food, and wine.

142

3GBX Vertical Studio

Instructor: John Enright

3GBX


2GAX

143


View From vineyard

144


145


Caleb Roberts

Structural Axon

Floorplan

146


3GBX

Section

East Elevation

South Elevation

147


Caleb Roberts

Interior View

148


3GBX

149


Caleb Roberts

SL-1 In 2011, an earthquake caused a meltdown at the Fukushima Daiichi nuclear power plant resulting in the spread of radioactive material and the creation of a 20km evacuation zone around the plant. In the years since, the environment within the evacuation zone has mutated due to the absence of humans and the ecological effects of radiation on plant and animal life. This project exaggerates these effects on the environment within virtual landscapes to explore changing aesthetics of nature. The features of these virtual landscapes traditionally viewed as beautiful are amplified and distorted to create a new aesthetics of beauty that comes out of what can only be described as an ecological disaster.

150


2GAX

151


Caleb Roberts

Virtual Landscapes 152


EDGE

153


Caleb Roberts

Virtual Landscapes 154


EDGE

155


Caleb Roberts

Virtual Landscape 3

156


2GAX

157


Caleb Roberts

SL-2 A vast majority of the landscape imagery available on the internet reproduces, often unintentionally, the outdated aesthetics of the Picturesque. This unconscious aesthetics ignores parts of the landscape touched and scarred by the human hand in an effort to capture true “nature”. However, my project critiques the aesthetic regime of the Picturesque, arguing that we cannot ignore these “unnatural” parts of the landscape. Utilizing image-based AI tools, my project contaminates datasets of picturesque landscapes with imagery of the ecological impacts that our society has had on the landscape. While the AI attempts to understand the collection of images and their similarities, it combines the two aesthetics into an agonistic relationship where the natural and the unnatural push against each other to emerge within the image.

158

Synthetic Landscapes Design Studio

Instructor: David Ruy

EDGE


2GAX

159


Caleb Roberts

Stylegan2 Dataset Images

160


2GAX

161


Caleb Roberts

Synthetic Landscape.01 -generated with Stylegan2

162


EDGE

Synthetic Landscape.25 -generated with Stylegan2

163


Caleb Roberts

10,000 Synthetic Landscapes -generated with Stylegan2

164


2GAX

165


Caleb Roberts

Rewilding Infrastructure

American infrastructural networks were created to connect cities, states, and humans to one another. While this succeeded in connecting our society together, it also carved through many nonhuman landscapes and destroyed numerous ecologies across the US. Nonhuman spaces were invaded by concrete and steal, forcing human aesthetics into these landscapes. Today, many highway networks lie unused or unfinished. Technological development is trending toward a future where these networks are no longer a necessary part of society. This project seeks to rewild the infrastructure of the US in order to bring together the human and nonhuman aesthetics of these spaces and make landscapes where the two cohabit instead of compete. Using AI style transfer, the aesthetics of the landscape surrounding three different infrastructural objects are projected back onto the forms, resulting in formal and material variations that are neither purely human nor purely nonhuman.

166

Third Chair Scenogrophes Seminar

Instructor: Tucker Van Leuwen-Hall

2GAX


2GAX

167


Caleb Roberts

Original Highway Interchange

Transformed Highway Interchange

168

Ecological Collage


EDGE

169


Caleb Roberts

Original Highway Overpass

Transformed Highway Overpass

170

Ecological Collage


EDGE

171


Caleb Roberts

Original Highway Overpass

Transformed Highway Overpass

172

Ecological Collage


EDGE

173


Caleb Roberts

174


2GAX

175


Caleb Roberts

AI Assemblies

AI Assemblies is a proof-of-concept for creating an automated workflow for producing low-cost ADU housing in Los Angeles. Leveraging the abilities of Machine Learning, we developed a process that begins with site selection and continues all the way through to roboticassembled timber structures. The goal of this workflow is an end-to-end design tool that can reduce costs for low-income housing in LA. The project begins with a Pix2Pix model trained to identify potential sites within satellite images of LA that meet the code requirements for ADUs. The model then generates footprints and floorplans for possible ADUs specific to that site. The floorplan is then used as an input into a script that develops a three-dimensional voxel structure that is broken down into structural elements suited for timber construction. The script takes into account stress vectors to efficiently organize the structural elements and prepare them for robotic assembly. Finally, panel schedules and shop drawings are created to be used in conjunction with robotic arms to assemble the ADUs on-site.

176

AI Assemblies Seminar

Casey Rehm

2GAX


2GAX

177


Caleb Roberts

Viable Lots

AI Generated Footprints

Extracted Footprints

178


EDGE

179


Caleb Roberts

Generated Floorplans

Generated Structures

Generated Structures

180


EDGE

181


Caleb Roberts

29'-9"

14'-8"

18'-1"

23'-4"

28'-0"

28'-0"

9'-11"

4'-8"

15'-1"

14'-6"

3'-2"

11'-11"

29'-7"

Floor Plan 1/4”=1’-0”

Floorplan

182


EDGE

Panel organization, structural type segmentation

183


Caleb Roberts

Timber Structure

184


EDGE

Birds eye 1/4”=1’-0”

185


Caleb Roberts

Ambiguous Forms

Ambiguous forms are created from the merging and morphing of multiple heterogeneous inputs. The Machine Learning model, Stylegan2, has the capacity to take collections of disparate images and find some common form within the dataset. The resulting images are not entirely a mixture or merging of the various inputs (A+B≠AB), but something wholly new (A+B=C). This project leverages the generative capabilities of Stylegan2 to explore a new aesthetics that emerges when multiple datasets collide. Datasets of Brutalist architecture, Yugoslavian war monuments, and Greek statues are combined to train a Stylegan2 model. Each of these was chosen for their use of material and for the formal language they use. The resulting imagery does not quite resemble any of the three inputs. It is a series of ambiguous forms.

186


2GAX

187


Caleb Roberts

Stylegan2 Output

188


2GAX

Stylegan2 Output

189


Caleb Roberts

Stylegan2 Output

190


2GAX

Stylegan2 Output

191


Caleb Roberts

1000 Stylegan2 Outputs

192


2GAX

193


Caleb Roberts

Arithmetic Imaging

This thesis is an exploration of arithmetic procedures of image production that I hope can map some ways out of the problems that come from an over-reliance on subjective judgments. Namely, something beyond judgments of “I like” vs “I don’t like”. If I am inevitably exercising nothing more than my own tastes in how I make images, how would I ever produce anything that is truly new? Further, aren’t my subjective judgments biased relative to my history and social standing? Might an impersonal procedure of imaging open up new possibilities for my imagination?

got rid of the ones I didn’t like.”) but also multiply and divide Once an image is created, I allow AI linguistic tools to classify and describe the image. An AI trained on Google’s image database produces a series of labels which I then use to divide the images into linguistic categories.

This method of production and classification leads into some more conceptual problems of contemporary imaging beyond the merely technical. First, this method of imaging relies on the impersonal multiplication of signifiers which The first part of this project is a deep dive into does not care so much about the particular proxy rendering techniques. In a nutshell, meaning of one object or its relationship to proxy rendering is a concept in current 3d others, but cares more about the proliferation rendering software that allows you to multiply of these signifiers. At any time, I can swap and scatter 3d elements throughout a scene one or more of the 3d elements with another, in extremely large quantities not possible with producing an image that is entirely different more traditional modeling techniques. You and equally valued within the system. Second, control the output by manipulating a series of the implementation of AI tools allows me to variables, swapping one 3d element for another, automate the classification of images. I no longer and multiplying many proxy systems together have to meticulously model and render an to produce an image. For the purposes of this image to produce a specific meaning; instead, I thesis, I have created a library of 3d elements operate as an image producing machine, while that are multiplied together using this system the AI interprets the images for meaning. This to produce a large number of images. The collaborative workflow between human and nonimages produced for this thesis are arbitrary and human rejects individual subjectivity in favor of without explicit intentionality. The primarily goal cultural impersonality. was to design a workflow that could not only add and subtract, (That is, “I made another image; I

194

MArch.II Thesis

Advisor: David Ruy

2GAX


2GAX

195


Caleb Roberts

Assets

196


MArch.II Thesis

197


Caleb Roberts

198


MArch.II Thesis

199


Caleb Roberts

200


2GAX

201


Caleb Roberts

AI Generated Description: An aerial photograph of the architecture ruins on a historic site. The picturesque building, which houses a large library, was built in the late 15th century, and has been reconstructed to make it a cultural centre. There are large reflection pools and gardens in the central courtyard. The main building is a terraced building with large windows, and the walls of the terrace are painted with traditional red and white accents. There are small statues of people on the terrace and the walls of the central courtyard.

202


MArch.II Thesis

The last part of this thesis extends the human-machine collaboration even further. The descriptions of the images were created by a GPT-2 text generating AI model. GPT-2 is a linguistic model trained on a vast library of texts ranging from Wikipedia to literature and scientific journals. It’s human collaborators give it a brief input which it then uses to write a full text continuing the input that it is given. For each of the images produced in my thesis, I allowed the GPT-2 model to write a description based on the labels generated by the image classifier. By allowing these text descriptions to bleed back into the image, I am creating a feedback loop whereby I post-process the images to more accurately fit the description given to them by the GPT-2 model.

Edited Image to match AI description: The image is post-processed to better fit the narrative description given to it by GPT-2. Elements added to this image include windows, statues, and red and white painted accents.

203


Caleb Roberts

204


2GAX

205



Caleb Roberts M. Arch II Southern California Institute of Architecture



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.