English

EAST: Early Action Prediction Sampling Strategy with Token Masking

Computer Vision and Pattern Recognition 2026-04-21 v1

Abstract

Early action prediction seeks to anticipate an action before it fully unfolds, but limited visual evidence makes this task especially challenging. We introduce EAST, a simple and efficient framework that enables a model to reason about incomplete observations. In our empirical study, we identify key components when training early action prediction models. Our key contribution is a randomized training strategy that samples a time step separating observed and unobserved video frames, enabling a single model to generalize seamlessly across all test-time observation ratios. We further show that joint learning on both observed and future (oracle) representations significantly boosts performance, even allowing an encoder-only model to excel. To improve scalability, we propose a token masking procedure that cuts memory usage in half and accelerates training by 2x with negligible accuracy loss. Combined with a forecasting decoder, EAST sets a new state of the art on NTU60, SSv2, and UCF101, surpassing previous best work by 10.1, 7.7, and 3.9 percentage points, respectively.

Keywords

Cite

@article{arxiv.2604.18367,
  title  = {EAST: Early Action Prediction Sampling Strategy with Token Masking},
  author = {Iva Sović and Ivan Martinović and Marin Oršić},
  journal= {arXiv preprint arXiv:2604.18367},
  year   = {2026}
}

Comments

Accepted at ICLR 2026

R2 v1 2026-07-01T12:18:32.457Z