ERA: Expert Retrieval and Assembly for Early Action Prediction

TL;DR

This paper introduces ERA, a novel module for early action prediction that retrieves specialized experts to distinguish highly similar actions using subtle differences, achieving state-of-the-art results.

Contribution

The paper proposes a new Expert Retrieval and Assembly (ERA) module and an Expert Learning Rate Optimization method for improved early action prediction.

Findings

Achieves state-of-the-art performance on four public datasets.

Effectively uses subtle differences for early action discrimination.

Expert learning rate optimization improves model training.

Abstract

Early action prediction aims to successfully predict the class label of an action before it is completely performed. This is a challenging task because the beginning stages of different actions can be very similar, with only minor subtle differences for discrimination. In this paper, we propose a novel Expert Retrieval and Assembly (ERA) module that retrieves and assembles a set of experts most specialized at using discriminative subtle differences, to distinguish an input sample from other highly similar samples. To encourage our model to effectively use subtle differences for early action prediction, we push experts to discriminate exclusively between samples that are highly similar, forcing these experts to learn to use subtle differences that exist between those samples. Additionally, we design an effective Expert Learning Rate Optimization method that balances the experts' optimization and leads to better performance. We evaluate our ERA module on four public action datasets and achieve state-of-the-art performance.