Selective Reinitialization Algorithms for Preventing Plasticity Loss in Artificial Neural Networks

The AI Seminar is a weekly meeting at the University of Alberta where researchers interested in artificial intelligence (AI) can share their research. Presenters include both local speakers from the University of Alberta and visitors from other institutions. Topics can be related in any way to artificial intelligence, from foundational theoretical work to innovative applications of AI techniques to new fields and problems.

In this seminar from Amii and the Department of Computing Science, J. Fernando Hernandez-Garcia demonstrates a phenomenon in which neural network systems lose their ability to learn when trained on non-stationary data. The phenomenon occurs across a wide variety of systems, posing a significant obstacle to the development of continual learning systems. This talk presents selective reinitialization algorithms as an approach for maintaining the ability to learn, offering a promising solution to this significant obstacle

Bio:
Fernando is a PhD candidate at the University of Alberta, supported by the Alberta Machine Intelligence Institute and advised by Richard S. Sutton. In the past, he received an undergraduate and a master's degree from the University of Alberta. Fernando's lifelong goal is to develop learning systems capable of acquiring knowledge and skills in a similar fashion to animals and humans.

Видео Selective Reinitialization Algorithms for Preventing Plasticity Loss in Artificial Neural Networks канала Amii