MixANT: Observation-dependent Memory Propagation for Stochastic Dense Action Anticipation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wasim, Syed Talal, Suleman, Hamid, Zatsarynna, Olga, Naseer, Muzammal, Gall, Juergen
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909787114962944
author Wasim, Syed Talal
Suleman, Hamid
Zatsarynna, Olga
Naseer, Muzammal
Gall, Juergen
author_facet Wasim, Syed Talal
Suleman, Hamid
Zatsarynna, Olga
Naseer, Muzammal
Gall, Juergen
contents We present MixANT, a novel architecture for stochastic long-term dense anticipation of human activities. While recent State Space Models (SSMs) like Mamba have shown promise through input-dependent selectivity on three key parameters, the critical forget-gate ($\textbf{A}$ matrix) controlling temporal memory remains static. We address this limitation by introducing a mixture of experts approach that dynamically selects contextually relevant $\textbf{A}$ matrices based on input features, enhancing representational capacity without sacrificing computational efficiency. Extensive experiments on the 50Salads, Breakfast, and Assembly101 datasets demonstrate that MixANT consistently outperforms state-of-the-art methods across all evaluation settings. Our results highlight the importance of input-dependent forget-gate mechanisms for reliable prediction of human behavior in diverse real-world scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2509_11394
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MixANT: Observation-dependent Memory Propagation for Stochastic Dense Action Anticipation
Wasim, Syed Talal
Suleman, Hamid
Zatsarynna, Olga
Naseer, Muzammal
Gall, Juergen
Computer Vision and Pattern Recognition
We present MixANT, a novel architecture for stochastic long-term dense anticipation of human activities. While recent State Space Models (SSMs) like Mamba have shown promise through input-dependent selectivity on three key parameters, the critical forget-gate ($\textbf{A}$ matrix) controlling temporal memory remains static. We address this limitation by introducing a mixture of experts approach that dynamically selects contextually relevant $\textbf{A}$ matrices based on input features, enhancing representational capacity without sacrificing computational efficiency. Extensive experiments on the 50Salads, Breakfast, and Assembly101 datasets demonstrate that MixANT consistently outperforms state-of-the-art methods across all evaluation settings. Our results highlight the importance of input-dependent forget-gate mechanisms for reliable prediction of human behavior in diverse real-world scenarios.
title MixANT: Observation-dependent Memory Propagation for Stochastic Dense Action Anticipation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.11394