Poster
in
Affinity Workshop: Blog Track Session 2
It's Time to Move On: Primacy Bias and Why It Helps to Forget
Matthew Kielo · Vladimir Lukin
Halle B #3
Abstract:
'The Primacy Bias in Deep Reinforcement Learning' (Nikishin et al. 2022) demonstrates how the first experiences of a deep learning model can cause catastrophic memorization and how this can be prevented. In this post we describe primacy bias, summarize the authors' key findings, and present a simple environment to experiment with primacy bias.
Live content is unavailable. Log in and register to view live content