![[2506.15190] Learning Task-Agnostic Motifs to Capture the Continuous Nature of Animal Behavior](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2506.15190] Learning Task-Agnostic Motifs to Capture the Continuous Nature of Animal Behavior
Summary
The paper presents a novel framework, Motif-based Continuous Dynamics (MCD), to model animal behavior by identifying continuous motor motifs, enhancing the understanding of behavior dynamics beyond traditional segmentation methods.
Why It Matters
This research is significant as it addresses the limitations of existing behavior segmentation techniques that oversimplify animal behavior. By introducing MCD, the study provides insights into the continuous nature of behavior, which can improve applications in robotics, animal cognition studies, and machine learning models focused on dynamic systems.
Key Takeaways
- MCD framework uncovers interpretable motor motifs as latent functions of behavior.
- It models behavioral dynamics as continuously evolving mixtures of identified motifs.
- The approach captures complex animal behaviors more accurately than traditional methods.
- MCD has been validated across multiple tasks, showcasing its versatility.
- This research advances the quantitative study of natural behaviors in animals.
Computer Science > Machine Learning arXiv:2506.15190 (cs) [Submitted on 18 Jun 2025 (v1), last revised 26 Feb 2026 (this version, v3)] Title:Learning Task-Agnostic Motifs to Capture the Continuous Nature of Animal Behavior Authors:Jiyi Wang, Jingyang Ke, Bo Dai, Anqi Wu View a PDF of the paper titled Learning Task-Agnostic Motifs to Capture the Continuous Nature of Animal Behavior, by Jiyi Wang and 3 other authors View PDF HTML (experimental) Abstract:Animals flexibly recombine a finite set of core motor motifs to meet diverse task demands, but existing behavior segmentation methods oversimplify this process by imposing discrete syllables under restrictive generative assumptions. To better capture the continuous structure of behavior generation, we introduce motif-based continuous dynamics (MCD) discovery, a framework that (1) uncovers interpretable motif sets as latent basis functions of behavior by leveraging representations of behavioral transition structure, and (2) models behavioral dynamics as continuously evolving mixtures of these motifs. We validate MCD on a multi-task gridworld, a labyrinth navigation task, and freely moving animal behavior. Across settings, it identifies reusable motif components, captures continuous compositional dynamics, and generates realistic trajectories beyond the capabilities of traditional discrete segmentation models. By providing a generative account of how complex animal behaviors emerge from dynamic combinations of fundamental motor...