Abstract

We have a visitor Jay McClelland, from 3-5 pm on 14th September. He will give an informal presentation, during which he will raise some issues about what he sees as strengths and weaknesses in today’s AI systems and how he sees them differing from humans. After the visit to CBL , he will also give the CAIS Lecture 2023 from 18:30 – 21:00.

About the speaker

James L. (Jay) McClelland is a Cognitive Scientist who has used neural network models to explore the mechanisms of human and machine intelligence for nearly 50 years. In the late 1970’s he introduced a neural network model capturing the dynamics of activation flow through a neural network. He then teamed up with David Rumelhart, the inventor of the learning algorithm that powers today’s neural-network based language models and many other machine learning systems. Together they produced the two-volume work Parallel Distributed Processing (MIT Press, 1986) that kindled the second wave of neural network research beginning in the mid 1980’s. McClelland led the creation the Center for the Neural Basis of Cognition at Carnegie Mellon University and the University of Pittsburgh in the 1990’s, then moved to Stanford University, where he led the creation of the Center for Mind, Brain and Computation in 2008. He is the Lucie Stern Professor in the Psychology Department at Stanford, and holds courtesy appointments in Computer Science and Linguistics, and he is currently a Consultant Research Scientist at Google DeepMind.