Augmented Creativity is a project of IaaC, Institute for Advanced Architecture of Catalonia developed at Master in Advanced Architecture in 2017 of the Academic Program by:
Student: Lalin Keyvan
Faculty: Luis Fraguada.

The aim is to investigate creative human notions through a computational modelling approach with a specific model, the long short term memory theory.

Till today, the dialogue with computers and designers basically consists of command or order relationship. Users asked questions and computers answered and obeyed. It is also known that the idea of a partnership between human/user and the computer is not something new, in fact, it flourished with inventions of the computer. However, due to developments in technology, this idea recently is becoming more real than ever. With this project, the intention is extending the idea of participatory design. Not just designing for people but design with people. Therefore with the understanding of “not every user can design for themselves, but some cases designers might not be enough as well”, it is important to question the possible impacts of computers on the design industry rather than just being a tool?

Computers can work in well-structured areas of problem-solving or they can be used as tools to help designers, artists, architects. But can they really perform a creative work or how they become more involved in the process?If they can then how we are going to interact with the computer physically?

Another important concern is the novelty paradox. According to Derek Partridge, Jon Rowe” the novelty paradox can be solved by reconstructing the memory. It is possible that elements involved in constructing a new idea have come from other ideas, but they might not be the ideas themselves. The new idea may be constructed by some of the same processing elements that built the other ones. In this way the new idea, though originating from earlier ones, is actually different from all of them.The idea is novel as a combination of elements.”

With the recent development of technology allows machines to have the ability to learn the creative behaviour with deep learning strategies which is a machine learning technique where it assigns weights to neural networks to abstract the data and trains a model.

With having a false consciousness or familiar material such as art or culture, machines can learn creative behaviour using deep learning techniques and this can be demonstrated with an interactive installation which will use sound as the key element of creative work.

For Augmented Creativity project, a library called Tensorflow developed by Google is used. Tensorflow is an open source library helps to create neural network system in project’s code which is written in python programming language. In the code, first, we set up a database, declared variables such as highest and lowest notes, the number of hidden layers and learning rate. Then we used Gibbs Sampling as a helper function to sample from the probability distribution. After trained the code and run the graph, results saved into a folder. Multiple experiments helped us to experiment and observe the results with different data sets, learning rates, different instruments and genres.

Following months, the research will pursue real-time feedback and the physical aspects of the installation.