Research map / Label-efficient and robust learning / Self-supervised and contrastive
LSTM Self-Supervision for Detailed Behavior Analysis
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Bootstrap-retrain the paw detector through three rounds | FCN initialized with optical-flow detections · Rat paw detection test set | accuracy: 40.2 -> 82.1 | Helped |
| Jointly train posture features through behavior-sequence learning (ablation) | Rat posture representation · Rat posture similarity test set | accuracy: 72 -> 85.6 | Helped |
| Learn sequence representations by distinguishing real from temporally permuted sequences | CNN-LSTM · Rat sequence similarity test set | accuracy: 64.0 -> 80.5 | Helped |
| Use temporal-order self-supervision rather than DTW-based sequence clustering (ablation) | LSTM sequence representation · Rat sequence similarity test set | accuracy: 64.0 -> 80.5 | Helped |
| Apply the self-supervised sequence-ordering approach | Pose representation · Olympic Sports | mean AUC: 0.79 -> 0.83 | Helped |
| Transfer the Olympic Sports-trained representation without Leeds fine-tuning | Pose representation · Leeds Sports | PCP: 43.5 -> 46.6 | Helped |
| Use paw trajectories instead of appearance-based behavior representations | MIL grasp classifier · Rat grasp impairment classification | classification performance | Hurt |
Closest papers by meaning
- Interpretable Self-Supervised Facial Micro-Expression Learning to Predict Cognitive State and Neurological Disorders (AAAI 2021 · similarity 0.86)
- PULSE: A Synchronized Five-Modality Dataset for Sensorimotor Coordination in Long-Horizon Daily Activities (NeurIPS 2026 · similarity 0.85)
- Neural-Behavioral Representation of Natural Whole-body Movement in Monkeys (NeurIPS 2026 · similarity 0.84)
- DynaMo: In-Domain Dynamics Pretraining for Visuo-Motor Control (NeurIPS 2024 · similarity 0.83)
- Learning Options for Compositional Motor Control with Adapter Banks (NeurIPS 2026 · similarity 0.83)
- SPINT: Spatial Permutation-Invariant Neural Transformer for Consistent Intracortical Motor Decoding (NeurIPS 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.