Learning recurrent representations for hierarchical behavior modeling

TL;DR

This paper introduces a recurrent neural network framework that models hierarchical animal behaviors from motion data, improving action detection and capturing high-level phenomena without supervision.

Contribution

It presents a novel generative recurrent network with lateral connections for hierarchical behavior modeling and demonstrates its effectiveness on animal and handwriting data.

Findings

Unsupervised motion prediction enhances action detection with limited labels.

The network captures high-level attributes like writer identity and fly gender.

Generated motion trajectories appear realistic and useful for qualitative evaluation.

Abstract

We propose a framework for detecting action patterns from motion sequences and modeling the sensory-motor relationship of animals, using a generative recurrent neural network. The network has a discriminative part (classifying actions) and a generative part (predicting motion), whose recurrent cells are laterally connected, allowing higher levels of the network to represent high level phenomena. We test our framework on two types of data, fruit fly behavior and online handwriting. Our results show that 1) taking advantage of unlabeled sequences, by predicting future motion, significantly improves action detection performance when training labels are scarce, 2) the network learns to represent high level phenomena such as writer identity and fly gender, without supervision, and 3) simulated motion trajectories, generated by treating motion prediction as input to the network, look realistic and may be used to qualitatively evaluate whether the model has learnt generative control rules.