Skills in Architectural Education a New Paradigm

March 2020

Architectural Learning Model in Context of Artificial Intelligence

Neerja Babbar*
Ph.D. Scholar under the guidance of
Dr. (Ar.) Prabhjot Kaur**, Director, IKGPTU Mohali, Campus II, Punjab, India
email id: * [email protected]
email id: ** [email protected]
Abstract
Automated Intelligence is changing the world at a fast pace, and newer domains of wisdom are being
explored with the collaborative repositories of information namely, cloud storage, Artificial Intelligence
and Robotics are some of the promising break-throughs of the digital revolution. The gap between Human
Intelligence (HI) and Artificial Intelligence (AI) is narrowing day by day. The new generation is at the
threshold of Wisdom Age, where everything is big data; and knowledge is out there available at the click
of a mouse. However, the drawing board based teaching methods continue to favour the traditional
Architectural Design studios, with the belief that peer group mechanism harnesses creativity. This
research endeavours to bring forth AI compatible pedagogical model in Architectural Design curriculum.
Keywords: Knowledge, Artificial Intelligence, Learning Model, Architectural Design Intelligence, AI
Integrated Digital Design Pedagogy

Introduction
The theories of human evolution weave around the events retained in the memory. As memory-retention
is classified as ‘knowledge’, the techniques to acquire knowledge are listed as ‘taxonomy’. The teaching
and learning method or pedagogy has become the cornerstone of knowledge acquisition process. In
earlier times, people memorised the lullabies, folk tales and stories and narrated them from generation
to generation. The younger lot were encouraged to explore their surroundings and master their skills,
using the information stored in their individual memories. As society evolved, people began to specialize
in different vocations. Learning took a new role with the mentors or teachers, taking the centre-stage in
a civilized set-up. Teachings were documented and texts were handed down to the pupils as part of their
personal baggage of lessons. The top- down pyramid emerged with various pedagogical models adaptable
to different mechanisms. However, the approach to training in Architectural Design continued as
internship for disciples working under a master builder in the latter’s studio until the establishment of
‘Bauhaus School’ by Walter Gropius and advocated as the traditional method of learning Architectural
Design.
This paper presents an introspective review of literature to find out whether the traditional tools of
creative logic and innovative problem- solving processes hold significance in the Artificial Intelligence
enabled environment This research progresses with an interpretation of design program and process;
effectiveness of modelling objectives and virtual collaborations; and the receptivity of senses and reflex
responses; as learning mechanism. An insight into the undergraduate Architectural Design studio learning
models has helped the author evolve an effective model vis-a-vis digital paradigm.

Proceedings of the SAENP

 Skills in Architectural Education a New Paradigm
March 2020

1. Architectural Design Learning Model

Teaching Design in studios is contextual in nature. The processes for introducing a design problem in a
studio are multi-directional and multi-dimensional too. The instructors may initiate an inquiry with
environmental concerns, imploring the students to look for harmonized or contrasting solutions. At other
times, they may lead the class into investigating the structural or the historical relevance thus aiming at
technological resonance with the location. Whatever be the method, the design goal is consistently
achieved. And the events follow a systematic logic as depicted below:

3D
Form Modelling
Iterating Finalisation
Statement solutions
Conceptual Fragmentation
Library visualization
Problem Study
Statement
Fig. 1. Algorithmic Design Process. (source: author)

1.1. Architectural Program

The investigation into the process of design inherits two-fold methodology- either scientific technology or
a creative pursuit. The comparable features of design between the process (the journey) and the product
(the destination) signify:

Table 1. Comparison of Design Process with Design Product

Design as a process Design as a product
 Rational- linear, logical and sequential  Creative- non-linear, reflective and
generative
 Iterative repetitions  Intuitive innovations
 Journey embedded with investigations  Novel outcome of creative instincts based
and structured interventions on knowledge of multiple skills and domains
 Simulated modelling of interim  Remodelled experiences reminiscent of
solutions practical applications
 Formal Analysis of probable solutions  Intuitive Analysis of probable acceptance
with respect to tangible constraints with respect to intangible user response
 The iterative journey leads to artistic  Aesthetically appealing and functionally
destination1. viable product
Architectural program is an offshoot of the thought process that develops overtime with the
fragmentation of the problem statement. Deshpande(Gadkari Uday: Deshpande Rahul : Bhonsle
Kirti:Nagpure Harshwardhan : Ganorkar Harshal, 2016) et. al, have reiterated the program as, “Design is

1
In the words of Frank Gehry, “The end-product rises to Art”.
Proceedings of the SAENP
 Skills in Architectural Education a New Paradigm
March 2020

regarded by many as the core problem-solving process of technological development. It is as fundamental

to technology as inquiry is to science and reading is to language.”
Christopher Benninger attributes good design to simple procedures and rational processes while
introducing the book - ‘An Introduction to PREDESIGN’ by Shah and Prasad (Babbar N, 2015). He stresses
that observation precedes innovation. And further supports his stand, with the premise that adopting a
sequential exercise for problem solving streamlines design goals. According to him, the logical process
makes the designers- considerate, articulate and smart. In this book, the distinction between
Predesign/context and the final design product has been illustrated as two sides of Architectural Program.
The initiatives are characterized with Space Relationship Diagram for site related information and
Functional Relationship Diagram for distribution of spaces, thereby, supporting the design criteria.

Problem Problem
Statement Solution

Architectural DESIGN
PREDESIGN Program

Analysis Synthesis

Rational Intuitive

Fig. 2. Scientific Approach to Architectural Program as illustrated in the book –

‘An Introduction to PREDESIGN’ by Shah and Prasad

1.2. Integrating Digital Intelligence with Architectural Design Intelligence

Architectural Design Intelligence is the domain of designers whose prerogative is to lead the designs as
algorithm of expression (sketches), integration (constraints and parameters) and organisation
(presentation and collaborations), or in other words- Architectural Program.
The arrival of Digitalization is embedded with the precondition, that it has to follow the demand for design
and not vice-versa. Sculptural shapes can be built with the help of coded components just the same way
as traditional designs were moulded out on formwork piece by piece thus beholding the Digital
Intelligence Model semiotic to Architectural Programming. Viz; Frank Gehry2 designs with CATIA(Alvarado
& Muñoz, 2012).
If imagination is synchronous with ideating solutions, then Lynn (Lynn & Kelly, 1999) argues that Architects
have always been familiar with thinking virtually, since they produce only drawings and not buildings.
Virtual Reality (VR) platform ensures the same product to the stake-holders as they see it (what you see
is what you get). The models generated on VR evoke reflex responses during the process of modelling
itself and the process gets an economic boost with all the collaborators working on the models
simultaneously. This coherence accounts for integrating Digital Intelligence with Architectural Design

2
Gehry designed ‘The Fish Barcelona’ in 1992 on CATIA developed by Dassault Systems
Proceedings of the SAENP
 Skills in Architectural Education a New Paradigm
March 2020

Intelligence. This integration gains a strong-hold because of the geometry shared between parametric
modelling and compositional design.
The repetitive patterns of fractal geometry, vector-based generators called Directrix, and techniques of
rapid prototyping ease the industrious digital fabrication processes. Manoeuvring hands to iterate
conceptual spatial models in virtual environments and digital sketching applications, help the designers in
controlling prospective propositions, contradictions and illusions. VR smoothens the optimization of the
design journey on the concepts of ‘Immersion, Interaction and User Involvement’. The users, who are
engaged in modifying and prototyping models in immersive media, are enriched with futuristic
experiences. The role of the architects extends beyond the boundaries of studio to include the engineering
analysis, energy optimization, project management and post-occupancy maintenance etc. And with
Augmented Reality (AR), the users are able to analyse sensory and visual experiments. Hence the inter-
operability of tools is of utmost importance which should facilitate communication between the
collaborators.

1.3. Facets of Learning

The learning mechanism has been classified into skill-based domains- Cognitive, Affective, Psycho-motor,
Relational (C.A.P.R.). While each learner associates with information uniquely, the way he translates it to
data stimulates different reflexes. The C.A.P.R. typology induces learning when both mind and body act in
coordination:

Table 2. Domains of learning (source: author)

Typology Application Stimulus
Cognitive (C): engaging mind and brain i.e. mental skills Visual
to acquire Knowledge
Affective (A): engaging mind and heart i.e. attitudinal skills Sensory and
to acquire Value-based learning auditory
Psycho-motor (P): engaging mind and body i.e. manual skills Tactile
to acquire Hands-on learning
Relational (R) engaging mind in peer-group activities Intrapersonal
i.e. communication and collaborative skills
to acquire social learning

Many a thinker has devised a systematized model of learning and updated it with the prevailing trends.
Vis; Benjamin Samuel Bloom3 proposed a taxonomy of learning goals based on C.A.P. model of learning.
While proposing digital vocabulary to Blooms taxonomy, Andrew Churches(Churches, 2008) referred to a
two-sided scale – Lower Order Thinking Skills or LOTs at the bottom of pyramid beginning with
‘Remember’ to Higher Order Thinking Skills or HOTs at the top of pyramid culminating in ‘Create’. His
intent was to target the students’ engagement with learning; right from lesson planning to course
mapping.

3
Benjamin Samuel Bloom , a doctorate in education from University of Chicago, edited the first volume of The
Taxonomy of Educational Objectives: The Classification of Educational Goals in 1956.
Proceedings of the SAENP
 Skills in Architectural Education a New Paradigm
March 2020

Create- formulate new solutions

Evaluate- assess program

Analyze components for
attributes
Apply- implement inferences to
futuristic exercises
Understand- infer interpretation of
stored data
Remember- recall knowledge from
memory

Fig. 3. Bloom’s Taxonomy in consonance with Digital Architecture Design (source: author)

Sabu Francis (Francis, 1989) states that computing is not just a tool but is the way to communicate. It is a
tool for programming4 and an epistemological device that helps stack (operational chatbots), heap (long-
term memory), provoke serendipity (coincident occurrences), do static and dynamic modelling or just-in-
time modelling, detect and utilize patterns.

2. An Insight into Artificial Intelligence

A machine manufactured to behave intelligently is referred to as Artificial Intelligence (AI). AI is
characterised by its ability to comprehend, analyse and assimilate knowledge. Whatever is stored in the
computing memory, be it images, information, mind-maps; AI can perceive it as multi-layer data base
(hitherto called, research) like site information; rationalize probabilities to further propose iterative
solutions in 3D modelling software tools (addressed as design); and clue up experiences for future
assignments. Thus, AI has the potential to facilitate the designers in selecting design variables, generating
alternatives and testing them for the most optimum selection. The final outcomes get registered in the
memory as aspects of AI enabled knowledge management.

Perceiving Rationalizing Planning Executing Learning

Fig. 5. AI enabled knowledge Management (source: author)

The skill- based domains can, hence, be rearranged in a reflexive model of Artificial Intelligence, as
illustrated in The Apttus5 Intelligence Capability Model(Shankar, 2018)

4
Programming is iterating models to achieve final product, thus discarding undesired ones.
5
Forbes list of top 100 cloud-based solutions recognizes the enterprise software company- ‘Apttus’ founded in
2006, as a leader.
Proceedings of the SAENP
 Skills in Architectural Education a New Paradigm
March 2020

Fig. 6. The Apttus Intelligence Capability Model (source: www.forbes.com/sites/louiscolumbus

/2016/04/04/the-era-of-the-intelligent-cloud-has-arrived/#349ff5971967)

2.1. AI Taxonomy
A Danish architectural practice, ‘3XN’(Edem Agbodjan, Kristian Prieme, Maja Olsson, Paul Jaeggi, 2019a)
and its subservient company ‘GXN’(Edem Agbodjan, Kristian Prieme, Maja Olsson, Paul Jaeggi, 2019b),
along with students of Master program in Strategic Design & Entrepreneurship, initiated an inquiry into
the role of AI in letting architects exploit research and strategic design. Their findings have been published
in the form of reports in June 2019. Edem et. al., quoted the research question as: “How could Artificial
Intelligence affect the creative design process within an architectural practice?” The objective of the
investigation was stated as identifying the possibilities for applying AI in Architecture for procuring
information, utilizing it to imbibe inferences and curating learning for future projects. The whole process
of handling a project right from its inception to finally handing it over to the clients, was classified into
three main activities: research, design and knowledge management, with all the teams as actors
collaborating in a network to come out with desired solutions.
Edem et. al. devised a taxonomical model pertaining to the learning domains; with generative design
(akella, 2016), natural language generation, natural language processing, machine learning, deep learning
and artificial neural networks stimulating the visual learning aspects of the cognitive domain; big data,
quantum computing, internet of things and robotics evoking AI-hardware based response; biometrics and
personalised treatments stimulating the sensorial facets of affective learning; artificial co-workers and
chatbots responding to tactile forms of psychomotor skills; self-driving vehicles and 3D mobility adding
to the kinesthetics of knowledge acquisition process. The reports confirmed that Machine learning6

6
Self-training by computing machines for applying statistical algorithms to process big data (for Architects, Data
comprises of sketches, technical drawings, models, regulations etc.) is referred to as Machine Learning., which is
further classified as:
i. Processing untagged images is called Unsupervised Learning;
ii. Processing labelled images is called Supervised Learning;
iii. Rewarding themselves for correct actions is Reinforced Learning;
iv. Neural networks imitating functioning of human brain to let information flow is called Deep Learning.
Proceedings of the SAENP
 Skills in Architectural Education a New Paradigm
March 2020

and Internet of Things (IOT)7 can capture information and restructure it to create distinctive AI
environment - parallel to the creative mind, yielding design solutions that are more people-oriented than
regulation-oriented.

2.2. Diagramming and Programming

When Architects embark on a project, they do a lot of case studies, literature studies, library studies and
site studies to get to the essence of the project. AI does the same with big data thus saving time on
understanding the nuances of the design exercise. AI is signified by the machines’ capability to harvest
experiential knowledge in new situations, rationalize its actions and also to analyse the outcomes as
learned experience. AI enriches learning acquisition process with DIKW model(Rosenblatt., 2014).

D I K W
•DATA - raw information captured •INFORMATION - •KNOWLEDGE - how to •WISDOM -when
from different sources contextually filtered data use filtered data and why to use it

Fig. 8. Rosenblatt’s DIKW model of Knowledge Acquisition (source: author)

The self-reliant form of AI can be entrusted to deliver responsive and intuitive solutions by bringing
together multiple technologies with varied attributes. Its capacity to transform tacit8 knowledge to
explicit9 knowledge without human intervention provokes a comparison with the human senses –
outcome similar to peer-group functionalities. Stancombe Christopher of Capgemini(Stancombe
Christopher, 2017), reiterates if human intelligence (HI) is a mix of senses, knowledge and experiences;
then AI has the potential to clue up to these senses.

a. General AI deals with all kinds of activities a human mind can perform
b. Narrow AI deals with specific activities like self-driving cars, smart speakers, web search etc.
7
Information captured with the help of sensors and other IOT instruments at site can be analyzed and controlled for
improving performativity.
8
Tacit knowledge is experience based knowledge embedded in our memory
9 Explicit knowledge is articulated knowledge embedded in rule-books or information systems.

Proceedings of the SAENP

 Skills in Architectural Education a New Paradigm
March 2020

Fig. 9. Five Senses of AI (source: www.capgemini.com/2017/05/the-five-senses-of-artificial-intelligence/#)

3. Architectural Design Intelligence Learning Model

Digital Intelligence (DI) integrates with Architectural Design Intelligence (ADI) to manoeuvre the design
program into a sequential journey. AI can be embedded into Architectural Design teaching methodology.
The five senses of AI – Act, Think, Remember, Watch, Listen/ Talk provoke the imagery- Kinesthetics,
Tactile, Memorial, Visual and Auditory /Speech respectively. The digital studio characterized by the
variables of immersion, interaction and user involvement, calls for infrastructural remodelling of
Architectural Design institutes. Immersive laboratories connected to cloud servers support peer-group
learning and enhance multi-disciplinary skills, that are highly expected of the Wisdom Age10. The term
‘DIGITAL’ can be seen as an acronym for ‘Define, Interpret, Grade, Iterate, Translate, Apply and Learn’.
Hence the DIGITAL model postulated in response to Digital Design Intelligence can be reproduced as
Architectural Design Intelligence Learning Model (ADILM):

DEFINE BRIEF
Observe
LEARN FEEDBACK INTERPRET/ RESEARCH
BACKGROUND
Remember
Listen

DIGITAL

APPLY/ IMPLEMENT GRADE/ IDEATE

DELIVERY SOLUTION
Simulate Think

TRANSLATE/ SELECT ITERATE/ PROTOTYPE

RATIONALE RESOLVE
Act Analyze

Fig. 10. Architectural Design Intelligence Learning Model (ADILM) (source: author)

4. AI Integrated Digital Design Pedagogy

The Architectural Design Intelligence Learning Model is synchronous with Digital Design pedagogy.
Arciszewski et.al. (Pour Rahimian, Arciszewski, & Goulding, 2014) have ascribed the pedagogical process
to the variables of course contents, instructors, physical environment and ambience. Design visualization
involves empathizing with a situation prior to conceptualizing design solutions. In order to yield solutions
that can be addressed as iterative/ generative/ morphogenic/ biomorphic or genomic; the cross-
disciplinary pedagogy for Computer Augmented Studio is recommended to include:

10
Wisdom Age follows 4th Industrial Revolution- an era leading to Artificial Intelligence (AI), Virtual Reality (VR) and
3D printing; while 3rd Industrial Revolution led to digital era - late 20th century; 2nd Industrial Revolution led to
telecommunication and information period - early 20th century; 1st Industrial Revolution led to Industrial mode-
late 18th century.
Proceedings of the SAENP
 Skills in Architectural Education a New Paradigm
March 2020

Table 3. Computer Augmented Studio Tools (source: author)

Course Tools Components
Arch. Visualization Communicative Social, Interactive, Coexisting, Curatorial
Game Design and Edutainment11/ Infotainment NURBS12, Fluid, Smooth, Liquid Geometry
Creative Indices
Virtual Design and Computational Techniques Scripting, Programming, Iterating, Rapid
Construction Prototyping,
Robotic Fabrication Software Components Rendering, Texturing, Networking, and AI

The exercise to remodel the Architectural Design institutes, requires a clear perception of the outcomes
in terms of knowledge domains and the students’ growth. Hence, the infra-structure compatible to
intelligence domains resounds with the establishment of Parametric Modelling Labs for harvesting
Analytical Intelligence; Computer Augmented Studios and Immersive Labs for empowering Innovative
Intelligence; Exhibition and Inventive Interaction Spaces for pushing the Exploratory Intelligence and the
Model Making and Building Mechatronics workshops for reaping Practical Intelligence. The ‘Transaction
of Learning’ promulgates distinct domains when the contexts for Architectural Design subscribe to the
languages of AI. With the result, a new model of teaching Architectural Design in AI enabled environment
emerges:

Fig. 11. AI Integrated Digital Design Pedagogy (source: author)

11
Edutainment is encouraging learning by assimilating the scale and shape grammar of objects
12
NURBS: Non Uniform Rational Basis Spline

Proceedings of the SAENP

 Skills in Architectural Education a New Paradigm
March 2020

5. Conclusion
The program of Architecture, is multi-disciplinary in nature. The cohesiveness, with which it binds the
parameters of Arts, Sciences, Humanities and Technology, urges the future Architects to stay ahead of the
prevailing industry standards because employability and adaptability to change comprise the scales to
measure the relevance of education. Such cause-effect relationship spurs the validation of edutainment13
as an eminent part of core curriculum and builds a climate of trust for Digital Design pedagogy.
Computers have ushered in ‘Intelligence Revolution’ and ‘Wisdom Age’ is a product of the same. Machines
are being designed to replicate humans so that humans could spare themselves from the labour of all the
repetitive /menial jobs and concentrate on more productive and creative ones. Although AI is not
contesting the empathetic nature of a creative mind yet it is powerful enough to present binomial
solutions. The Architectural Design Intelligence Learning Model and AI Integrated Digital Design Pedagogy
are developed to optimize studio pedagogy with the aim to instill collaborative skills in young minds.

References:
Akella, R. (2016). What Generative Design Is and Why It’s the Future of Manufacturing. Retrieved from Industry
Week website: https://www.industryweek.com/technology-and-iiot/what-generative-design-and-why-its-
future-manufacturing
Alvarado, R. G., & Muñoz, J. J. (2012). The control of shape: Origins of parametric design in architecture in xenakis,
gehry and grimshaw. Metu Journal of the Faculty of Architecture.
https://doi.org/10.4305/METU.JFA.2012.1.6
Babbar N. (2015). Review of book-’An Introduction to Pre-Design’by Shah Naresh with Anaokar Prasad: NIASA,
Council of Architecture, 2015; 120 pages. ISBN No. 978-93-84204-02-0. International Journal of Habitat
Design & Environment, 1(II), 81–83.
Churches, A. (2008). Bloom ’ s Digital Taxonomy. Thinking, (July), 1–44. https://doi.org/10.1002/ejoc.201200111
Edem Agbodjan, Kristian Prieme, Maja Olsson, Paul Jaeggi, S. S. M. (2019a). Artificial Intelligence in Architecture.
Edem Agbodjan, Kristian Prieme, Maja Olsson, Paul Jaeggi, S. S. M. (2019b). GXN report on AI for Architecture.
Retrieved from https://linkinghub.elsevier.com/retrieve/pii/S1524904206000063
Francis, S. (1989). TAD A unique BIM that is for the rest of us (p. 1). p. 1. Retrieved from www.teamtad.com
Gadkari Uday: Deshpande Rahul : Bhonsle Kirti:Nagpure Harshwardhan : Ganorkar Harshal. (2016). Learning by
Design. International Journal of Habitat Design & Environment, 2(I), 22–30.
Lynn, G., & Kelly, T. (1999). Animate form. In Animate Form. Retrieved from
http://www.annapujadas.cat/CSIM/estetica/textos/animate_form.pdf
Pour Rahimian, F., Arciszewski, T., & Goulding, J. S. (2014). Successful education for AEC professionals: case study
of applying immersive game-like virtual reality interfaces. Visualization in Engineering, 2(1).
https://doi.org/10.1186/2213-7459-2-4
Rosenblatt., G. (2014). Knowledge and Artificial Intelligence When Machines Know. Retrieved from www.the-vital-
edge.com/knowledge-and-artificial-intelligence/
Shankar. (2018). Web personalisation with Drupal and Machine Learning | Opensense Labs. Open Sense Labs.
Retrieved from https://opensenselabs.com/blog/articles/machine-learning-drupal
Stancombe Christopher. (2017). The five senses of Artificial Intelligence. Retrieved from capegemini.com website:
www.capgemini.com/2017/05/the-five-senses-of-artificial-intelligence/#

13
Combination of education and entertainment
Proceedings of the SAENP