# Shaped by What's Missing: Topological Invariance Simplification Discovers Behavioral Transitions

DevFeed: [Shaped by What's Missing: Topological Invariance Simplification Discovers Behavioral Transitions](<https://devfeed.tech/articles/shaped-by-what-s-missing-topological-invariance-simplification-discovers-behavioral-transitions-114433.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.00688>)

Author: Kaiwen Bian, Eric J. Leonardis, Yuanjia Yang, Charles Zhang, Yusu Wang, Talmo D. Pereira

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [clustering](<https://devfeed.tech/topics/clustering.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [graph theory](<https://devfeed.tech/topics/graph-theory.md>), [Sports](<https://devfeed.tech/topics/sports.md>)

Tags: [animals](<https://devfeed.tech/tags/animals.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [clustering](<https://devfeed.tech/tags/clustering.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cycles](<https://devfeed.tech/tags/cycles.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [graph](<https://devfeed.tech/tags/graph.md>), [invariance](<https://devfeed.tech/tags/invariance.md>), [network](<https://devfeed.tech/tags/network.md>), [neural](<https://devfeed.tech/tags/neural.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [physics](<https://devfeed.tech/tags/physics.md>), [rat](<https://devfeed.tech/tags/rat.md>), [space](<https://devfeed.tech/tags/space.md>), [spaces](<https://devfeed.tech/tags/spaces.md>), [state](<https://devfeed.tech/tags/state.md>), [structure](<https://devfeed.tech/tags/structure.md>)

## AI overview

The method identifies behavioral transitions directly from trajectory structure, without first segmenting time into behavioral states. It treats sustained behaviors as dense regions in a feature-space density landscape and transitions as sparse connecting corridors or loops. From frame-by-frame trajectories, it builds a discrete-Morse graph, extracts paths and cycles, and clusters them as transitions. On rat behavior data across three feature spaces--including a neural-network controller's latent space--the method achieved coverage comparable to state-space and topological-clustering baselines, while encoding more transition-identity information, producing clusters at least as tight, and matching annotations with less reliance on time warping. The abstract reports that transitions were consistent across the tested feature spaces.

## Source excerpt

arXiv:2611.00688v1 Announce Type: new Abstract: Understanding how animals move between behaviors is a central question in neuroethology, yet most segmentation methods first partition time into discrete states, leaving transitions as the unmodeled residue between them. We take the opposite view: in a suitable feature space, sustained behaviors are dense regions of the trajectory's density landscape, and transitions are the sparse corridors and loops that connect them. Because this structure is a property of the landscape's shape rather than of any particular coordinates, it can be read off directly with topology and should persist across feature spaces. We build this landscape from frame-by-frame trajectories, extract its discrete-Morse graph, and keep two kinds of structure: paths, the corridors a trajectory follows between dense behavioral regions, and cycles, closed routes whose entry and exit take different ways. Clustering these structures yields transitions directly, with no upstream state partition. On a rat behavior dataset across three feature spaces, including the latent space of a neural-network controller that imitates the animal in a physics simulator, the approach recovers transitions at coverage comparable to state-space and topological-clustering baselines, while carrying substantially more information about the transition identity, forming clusters at least as tight as the baselines', and matching annotated transitions with less reliance on time-warping. More broadly, the results show that behavioral transitions are a topological feature of a representation: recoverable from the shape of its trajectories, and consistent across distinct feature spaces.