The study of artificial intelligence (AI) and neuroscience has many things in common. At its core, neuroscience aims to better understand the brain by deciphering its complex networks and processes.
Complementarily, many AI-focused research projects involve building synthetic components of the human brain. Connecting these fields benefits both computer scientists and biologically-focused neuroscientists because they help us understand natural and artificial learning systems.
These areas of research lend themselves to drawing inspiration from each other. By the 1950s, researchers had already begun to see how they could model the information-processing abilities of human neurons. Today, AI is giving way to new tools for neuroscience research, some of which are contributing to new hypotheses about how cognitive processes and tasks are performed in the brain.
On Saturday, July 9, from 3:15 p.m. to 4:45 p.m., international experts will meet to discuss a wide range of topics related to the interplay between neuroscience and AI.
Dr. Kanaka Rajan, an assistant professor at the Friedman Brain Institute at the Icahn School of Medicine at Mount Sinai, takes data from neuroscience experiments and combines it with powerful computational frameworks to create artificial models of the brain. As a computational neuroscientist, his work aims to bridge the gap between the desire of AI researchers to find high-performance systems for a specific purpose or application, and the goal of biologists to discover how a system solves problems. and to make predictions from models that can further drive new hypotheses about brain function.
During the session, Dr. Rajan will discuss curricular learning, a novel approach taken by her lab to probe learning mechanisms in biological and artificial brains. This approach mimics the order of meaningful learning in human programs by training a machine learning model starting with an increasingly difficult training set, or “programs”.
As an experimental cognitive neuroscientist, Professor Stanislas Dehaene of the College de France, will argue that our species is truly unique in its ability to learn and that at least for now our brains can still learn better than any machine. who exist !
In addition to elaborating on some of the work featured in his 2020 book, how we learn, he will also talk about new research carried out in his laboratory. One specific area he and his team are looking at is humans’ impressive ability to find patterns in sequences (as in grammar) or in space (as in geometry). The data the team has generated so far poses a challenge to current artificial neural networks, which have not achieved similar performance so far and are generally poor in symbols and grammar.
Federation of European Neuroscience Societies