Sofian Audry

In my work, I’m interested in hijacking and subverting computer science and A.I. technologies and sort of use them to reveal some aspects of Indian of I guess, of what makes us human. AI Most technologies are designed to be powerful, to control. Like Susan mentioned earlier, to enslave. I really like that idea, to enslave nature, to enslave us all and to to the same time, to maybe solve some of our problems, to cope up with our deficiencies and our imperfections. In my work, I’m interested in flipping that and revealing these very imperfections that I think makes makes us what we are. And I think that these these things can be found in these technologies that we we designed. So so most of my work revolves around these notions of agent and behavior agents being also coming from this computer science notion of agent based model comes from cybernetics. An agent being a an entity that acts in the world in response to observations and behavior of being like how we can how another entity would perceive this agent. So as an unchanging form of event due to the activity within an assembly. This come from Gordon Pass. Pask has a very broad notion of what an assembly is. He talks about the behavior of statue. I would personally I would argue on that, but so I would I would personally replace the word assembly with agent. But I’m interested in how we experience these behaviors now can be can be used to generate new experiences. So this is an example of a work of one of the first actually works where I used started using machine learning in 2011. It is a robot that was hanging to a cliff in the the catalog kept the Catalonian. And this robot is looking at the cliff right now because it it it doesn’t like sun. It doesn’t like to see the light. At the same time. It needs the light because it needs to recharge the batteries with the solar panels. So it needs to, you know, once in a while, once in a while look at the light. I gave it a I used a reinforcement learning system where to to to let this agent try to solve this conundrum over the summer. So I’m going to throw a few concepts here just to get the conversation ready. But I mean, like Ben already introduced machine learning, but I’m just going to give like what you know, I think what for me is a very good summary of to try to understand what machine learning is in comparison to artificial intelligence, which is so. Machine learning is a subset is a subfield of artificial intelligence. And artificial intelligence being this big project of trying to reproduce human performance in all kinds of spheres. And the problem is like is the following how can we get the probe? How can we program machines to do things we do when we ourselves do not know how we do these things? This is almost everything when we talk, when we walk, when we make decisions, most of the time we use our intuitions. We do these things without even thinking about them. So if we don’t know how we do them, how could we even program them into a machine? And the intuition here are from the basis of base of machine learning is, well, we don’t know how we do these things, but what we know is that we once did not know how to do these things and we learn how to do them by subjecting ourselves to experience and by adapting.

Machine learning has again taken that very interesting idea of this, an agent adapting to the real world. And and once again, I’m going to reuse that. I really like this concept of that that that Susan was talking about earlier. Enslavement, enslavement. So so machine learning has enslaved these little agents into doing something, something where the the agency of the agent doesn’t matter anymore. What we’re after is to solve a problem. So using a model, we train an agent on some data and we try to adjust it, make it adjust itself to to solve whatever problem we have, whether it’s the it’s going to be like speech recognition or recognizing faces or or making as much money as much money as possible into the stock market. I’m so this is a project I’m working right now. We’re I’m, I’m, I’m, I’m following kind of this similar idea I’m trying to unfold this process that of, of, of learning. So, I mean, like nowadays, if you look at machine learning algorithms, we use them every day. We don’t necessarily need to know how they work. But I think that for me, what’s interesting is how they work, how they do this learning, because I think there are some very interesting I see that there’s a potential there in looking at the system, learning, there’s a potential for revealing, again, some things about ourselves. This is a new project I’m working on right now. Going more focuses with was that it’s an ongoing you and Josette is a visual artist really interested in artificial life and and embodiment. And so we’re creating this we’re working on this kind of ball robot that will have a strange non continuous body and we’ll have to try to, to learn how to roll.