top of page

Artificial Intelligence and Machine Learning in Architecture

In the ever changing, fast paced world of today, artificial intelligence and machine learning are scientific advancements drawing the attention of learned experts and industry leaders alike; and not without reason. As if ripped from the pages of the science fiction novels of old, such technologies seek to mimic human intelligence and behavior, and they can be used in a wide variety of fields, including architecture.


Artificial Intelligence is a technology which also aims to replicate how a human brain thinks and give the best possible results by analyzing tons of data with fast iterative processing and intelligent algorithms, allowing the software to learn automatically from patterns or features in the data. It is all about how our brains create patterns that merge and join data to construct a version of the world within which we make decisions. So, data and data analytics are the core of any AI application.


The ability to utilize tons of previous data in a millisecond to enhance the architecture design process could work wonders. AI can transform the way we work and live, and can achieve city’s wellbeing, sustainability, and economic goals. It is a high technology mechanical system that can perform any task but need a few human efforts like visual interpretation or design-making etc.


AI can be used in simple daily life activities like recommending films to watch to more complicated ones like agriculture, medicine, architecture, and linguistics. As architects, it is important to focus on how AI could help us regarding the design process and other important topics. Machine learning - which is a subset of AI - can change the way we think of design and architecture and help us to save much time and effort.


Machine learning can mimic human intelligence. It is defined by scientist Tom Mitchel as

"The study of computer algorithms that allow computer programs to automatically improve through experience”. However, as architects and experts, how can it be implemented in our designs and change the way we think? In design and architecture context, it has great potential in analyzing designs more quickly and evaluating the performance of a floor plan, helping architects understand how well a floorplate works in terms of visual navigation, walkability, the balance of private offices and shared spaces. With AI’s ability to take limitless amounts of data, an architect could easily go about researching and testing several ideas at the same time easily.



Architecture studios are now trying to develop and implement machine-learning ideas. Norman Foster and partners, which is a leading architecture studio represented by the Applied Research and Development group, are trying to think smartly by not seeking to replicate or replace designers, but to enhance knowledge and sensitivities, alongside freeing us from mundane, repetitive tasks. To have a clear idea of how machine learning can help, let us take an example of the building’s typology; suppose that we want to identify and classify Palladian villas from among a list of other images. At first, we identify whether each building matches the building typology using machine learning. We can do this by feeding the network of thousands of images of villas and train it to choose the right villas that represent the Palladian style design. In this way, we are training it by experience how our Palladian villas can be identified.



Identifying building typologies was done by training the system to classify the villas thanks to the wealth of related images that can be fed to it, but how could the system be trained to answer subjective questions, those which are related to, say, characteristics, for example?

For example, if we want the system to choose successful designs for prominent architecture in public locations, we can feed the system with images of a variety of designs, both successful and unsuccessful, and point out the designs that possess successful characteristics.


Machine learning is still an emerging topic and we do not yet have all the needed answers. However, despite a few drawbacks and nagging doubts, it is quite clear that this burgeoning field of science will lead to a breakthrough in numerous other fields, including architecture, and the next generation of architects will have to harness the knowledge and capabilities of this new technology should they wish to reap the immense benefits and returns.

Comments


bottom of page