Pretrained Vision-Language-Action Models are Surprisingly Resistant to Forgetting in Continual Learning

TL;DR

Pretrained vision-language-action models exhibit remarkable resistance to forgetting in continual learning, with simple experience replay effectively preserving learned skills and large-scale pretraining fundamentally enhancing continual learning capabilities.

Contribution

This work demonstrates that large-scale pretrained VLAs are inherently more resistant to forgetting, and simple experience replay suffices for continual learning, highlighting the impact of pretraining.

Findings

Pretrained VLAs show minimal forgetting compared to smaller models.

Simple experience replay achieves near-zero forgetting on VLAs.

Pretraining enables rapid skill recovery after learning new tasks.

Abstract

Continual learning is a long-standing challenge in robot policy learning, where a policy must acquire new skills over time without catastrophically forgetting previously learned ones. While prior work has extensively studied continual learning in relatively small behavior cloning (BC) policy models trained from scratch, its behavior in modern large-scale pretrained Vision-Language-Action (VLA) models remains underexplored. In this work, we found that pretrained VLAs are remarkably resistant to forgetting compared with smaller policy models trained from scratch. Simple Experience Replay (ER) works surprisingly well on VLAs, sometimes achieving zero forgetting even with a small replay data size. Our analysis reveals that pretraining plays a critical role in downstream continual learning performance: large pretrained models mitigate forgetting with a small replay buffer size while maintaining strong forward learning capabilities. Furthermore, we found that VLAs can retain relevant knowledge from prior tasks despite performance degradation during learning new tasks. This knowledge retention enables rapid recovery of seemingly forgotten skills through finetuning. Together, these insights imply that large-scale pretraining fundamentally changes the dynamics of continual learning, enabling models to continually acquire new skills over time with simple replay. Code and more information can be found at https://continual-vlas.github.io/forget-me-not/