USTEP: Spatio-Temporal Predictive Learning under A Unified View

TL;DR

USTEP introduces a unified framework for spatio-temporal predictive learning that combines recurrent-based and recurrent-free methods, effectively capturing both micro and macro temporal dependencies to improve performance across various applications.

Contribution

The paper proposes USTEP, a novel unified framework that integrates two dominant temporal modeling approaches for enhanced spatio-temporal predictive learning.

Findings

USTEP outperforms existing methods on multiple benchmarks.

The unified approach captures both short-term and long-term temporal dependencies.

Experimental results demonstrate significant performance improvements.

Abstract

Spatio-temporal predictive learning plays a crucial role in self-supervised learning, with wide-ranging applications across a diverse range of fields. Previous approaches for temporal modeling fall into two categories: recurrent-based and recurrent-free methods. The former, while meticulously processing frames one by one, neglect short-term spatio-temporal information redundancies, leading to inefficiencies. The latter naively stack frames sequentially, overlooking the inherent temporal dependencies. In this paper, we re-examine the two dominant temporal modeling approaches within the realm of spatio-temporal predictive learning, offering a unified perspective. Building upon this analysis, we introduce USTEP (Unified Spatio-TEmporal Predictive learning), an innovative framework that reconciles the recurrent-based and recurrent-free methods by integrating both micro-temporal and macro-temporal scales. Extensive experiments on a wide range of spatio-temporal predictive learning demonstrate that USTEP achieves significant improvements over existing temporal modeling approaches, thereby establishing it as a robust solution for a wide range of spatio-temporal applications.