Decoupled Prompt-Adapter Tuning for Continual Activity Recognition

TL;DR

This paper introduces DPAT, a novel framework combining adapters and learnable prompts to enable continual action recognition, maintaining performance across evolving video data without extensive finetuning.

Contribution

We propose Decoupled Prompt-Adapter Tuning (DPAT), a new method that effectively balances generalization and plasticity for continual action recognition in dynamic video environments.

Findings

Achieves state-of-the-art results on multiple benchmarks.

Effectively mitigates catastrophic forgetting in continual learning.

Balances generalization and plasticity in pretrained models.

Abstract

Action recognition technology plays a vital role in enhancing security through surveillance systems, enabling better patient monitoring in healthcare, providing in-depth performance analysis in sports, and facilitating seamless human-AI collaboration in domains such as manufacturing and assistive technologies. The dynamic nature of data in these areas underscores the need for models that can continuously adapt to new video data without losing previously acquired knowledge, highlighting the critical role of advanced continual action recognition. To address these challenges, we propose Decoupled Prompt-Adapter Tuning (DPAT), a novel framework that integrates adapters for capturing spatial-temporal information and learnable prompts for mitigating catastrophic forgetting through a decoupled training strategy. DPAT uniquely balances the generalization benefits of prompt tuning with the plasticity provided by adapters in pretrained vision models, effectively addressing the challenge of maintaining model performance amidst continuous data evolution without necessitating extensive finetuning. DPAT consistently achieves state-of-the-art performance across several challenging action recognition benchmarks, thus demonstrating the effectiveness of our model in the domain of continual action recognition.