Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environment

Hosted by the Wellcome Neural Dynamics PhD Programme

A key challenge for AI is to build systems that must adapt to changing task contexts and learn continuously. Although standard deep learning systems achieve state of the art results on static benchmarks, they often struggle in dynamic scenarios where the task changes over time, and this phenomenon is known as catastrophic forgetting. In this talk, I will discuss how biophysical properties of dendrites in the brain and local inhibitory systems enable networks to dynamically restrict and route information in a context-specific manner. Specifically, I will highlight the performance of a deep learning architecture that embodies properties of dendrites on two separate benchmarks requiring task-based adaptation: Meta-World (a multi-task reinforcement learning environment where a robotic agent must learn to solve a variety of manipulation tasks simultaneously) and permutedMNIST (a continual learning benchmark in which the model's prediction task changes throughout training). Analysis on both benchmarks demonstrates the emergence of overlapping but distinct and sparse subnetworks, allowing the system to fluidly learn multiple tasks with minimal forgetting. This work sheds light on how biological properties of neurons can inform deep learning systems to address dynamic scenarios that are typically impossible for traditional ANNs to solve.

Contact information

Contact Luke Burguete with any enquiries.