Open Resources to Become Knowledgeable in the Field of AI – ArchDaily

Open Resources to Become Knowledgeable in the Field of AI

Facebook

Twitter

Pinterest

Whatsapp

Mail

Or

As Artificial Intelligence has become one of the most significant forces driving innovation and economic development, this societal transformation requires new knowledge and an additional set of skills. Just as knowing a BIM software has become a prerequisite for most architecture jobs, understanding or even knowing how to use AI-related tools would become a desirable asset, if not a requirement in the future. However, with a vast array of information available, how does one begin to venture into this topic? The following is a compilation of online resources, lectures, and courses, that could provide a better understanding of the field and how to incorporate it into the practice of architecture.

What does Artificial Intelligence represent, what is the difference between machine learning and deep learning? These notions might seem interchangeable and so navigating the topic could become confusing. Before diving into the actual list of resources, it is essential to have the proper use of the most common terms.

Artificial intelligence (AI) is a wide-ranging branch of computer science concerned with the development of systems able to perform tasks typically requiring human intelligence. The AI encountered in different applications today is Artificial Narrow Intelligence (ANI), or "weak AI", used on performing a specific task, within a limited context, following pre-programmed rules. Google search, personal assistants, image recognition software, all fall into this category. Artificial General Intelligence (AGI) or Strong AI is still the realm of science fiction, as it would entail the general intelligence of a human being, able to solve any problem.

Put simply, Machine Learning is a subfield of AI, which consists of feeding data to a computer and using statistics and trial and error to help the network learn how to get better at a task, without having been programmed explicitly for that task, thus eliminating the need for writing overwhelmingly extensive code. Machine learning allows computers to make connections, discover patterns and make predictions based on what they learned in the past. A great way of understanding how this works in practice is the visual introduction in machine learning, created by R2D3, which uses a hypothetical example to explain the machine learning process.

Deep learning is a type of machine learning that feeds the data through a neural network architecture inspired by the human way of processing information, known as Artificial Neural Networks (ANNs). An example of usage for machine learning and deep learning is Google Image search.

Generative Design is a buzzword that has penetrated the architecture field a while now (see Archdaily's coverage of the topic here), but can it be framed as Artificial Intelligence or is it just a problem solver engaging multiple variables? Generative design is an iterative and exploratory process, where the input consists of parameters such as spatial requirements, performance, material constraints, as well as design goals. The software explores all possible solutions. Whether it falls in the realm of AI or not depends on whether the software is capable of testing and learning from each iteration, thus "learning" to give optimized answers.

With the ambitious goal to educate 1 % of European citizens in the fundamentals of AI by 2021, Elements of AI, a series of online courses created by Reaktor and the University of Helsinki sets the foundation for understanding the field, explaining what AI is, what it can and can't do, and how to start employing AI methods. The course is free and available in multiple languages, the aim being to teach people from a variety of backgrounds on the basic concepts of artificial intelligence technology. With almost 400.000 students so far from over 170 countries, the course is indeed proving an accessible and engaging resource.

Another beginner course, Coursera's Introduction to Artificial Intelligence, is also a great place to start building up the foundation concepts of the field, and it also contains some hands-on exercises.

For a non-aficionado, navigating literature on the subject of AI can be daunting. Therefore this Machine Learning Glossary provided by Google is a fast and reliable way of checking the meaning of terms when facing specialized jargon. The concepts are explained in a clear, straightforward fashion, and the glossary is an information resource in itself.

Learning is always more successful with a hands-on approach, and you can get acquainted with AI tools without having to learn to code. Project Runway ML is a public beta software and a platform dedicated to creators of all kinds that allow them to use AI tools without necessitating coding experience. From object detection to generating images from sketches, or creating text descriptions for images, the platform is a fun way to explore some design applications of AI.

This discussion at Columbia GSAPP explores artificial intelligence in architecture through the lens of several research projects.

Harvard GSD's lecture presents how AI-based tools and computer simulations could support landscape architecture.

In this lecture at the Strelka Institute, sociologist and professor Benjamin Bratton talks about AI and shares the results of the research projects undertaken in collaboration with Google Research. Read more about Bratton's ideas concerning in this Archdaily interview.

In addition, you can now take a virtual tour of the AI & Architecture exhibition, which was scheduled to take place at the Pavillon de l'Arsenal in Paris, France, but closed down due to the COVID-19 crisis. The curators decided to offer to the public an immersive experience, by recreating the exhibit as a virtual tour. Featuring the opening conference, a timeline of AI development, examples of its application to architecture, the exhibition is very rich in information and indeed an immersive experience.

Going further would require maths, as well as computer science prior training. Still, there are plenty of online resources addressing a more knowledgeable audience.

Google's Machine Learning Crash Course does not require any prior knowledge in machine learning, but students should have some experience programming in Python. However, all the different topics have an Introduction page that can at least provide an idea about how the process works.

Another Machine Learning course, this time offered by Stanford University, focuses on gaining the practical know-how on the subject.

For those already skilled in computer science, there is also the Artificial Intelligence course available on MIT's Youtube channel.

For a more comprehensive picture of AI in architecture, see Archdaily's coverage of the topic here.

Read the original here:
Open Resources to Become Knowledgeable in the Field of AI - ArchDaily