Research map / Label-efficient and robust learning / Self-supervised and contrastive
Drop, Swap, and Generate: A Self-Supervised Approach for Generating Neural Activity
Ran Liu, Mehdi Azabou, Max Dabagia, Chi-Heng Lin, Mohammad Gheshlaghi Azar, Keith B Hengen, Michal Valko, Eva L Dyer
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Use Swap-VAE rather than MYOW | reach direction decoding · Chewie-1 | acc (%): 70.41 -> 72.81 | Improved |
| Use Swap-VAE rather than Sup-Reach | reach direction decoding · Chewie-2 | acc (%): 69.71 -> 68.97 | Hurt |
| Use Swap-VAE rather than MYOW | reach direction decoding · Mihi-1 | acc (%): 70.48 -> 64.26 | Hurt |
| Use Swap-VAE rather than MYOW | reach direction decoding · Mihi-2 | acc (%): 64.35 -> 66.12 | Improved |
| Use Swap-VAE rather than the strongest listed temporal-decoding baseline | temporal decoding · Chewie-1 | accuracy (%): 53.80 -> 56.09 | Improved |
| Use Swap-VAE rather than beta-VAE | temporal decoding · Chewie-2 | accuracy (%): 63.76 -> 62.54 | Hurt |
| Use Swap-VAE rather than Sup-Time | temporal decoding · Mihi-1 | accuracy (%): 55.93 -> 57.77 | Improved |
| Use Swap-VAE rather than Sup-Time | temporal decoding · Mihi-2 | accuracy (%): 58.75 -> 49.65 | Hurt |
| Add BlockSwap to the proposed loss (ablation) | Swap-VAE · Chewie-1 | acc (%): 63.17 -> 72.81 | Improved |
| Add the L2 alignment term alongside BlockSwap (ablation) | Swap-VAE · Chewie-1 | acc (%): 71.63 -> 72.81 | Improved |
| Use both spatial and temporal augmentations instead of spatial-only augmentation (ablation) | Swap-VAE · Chewie-1 | acc (%): 66.77 -> 72.81 | Improved |
| Use both spatial and temporal augmentations instead of temporal-only augmentation (ablation) | Swap-VAE · Chewie-1 | acc (%): 71.15 -> 72.81 | Improved |
Closest papers by meaning
- Exploring Behavior-Relevant and Disentangled Neural Dynamics with Generative Diffusion Models (NeurIPS 2024 · similarity 0.88)
- BLEND: Behavior-guided Neural Population Dynamics Modeling via Privileged Knowledge Distillation (ICLR 2025 · similarity 0.86)
- Neuroformer: Multimodal and Multitask Generative Pretraining for Brain Data (ICLR 2024 · similarity 0.86)
- Disentangling Representations through Multi-task Learning (ICLR 2025 · similarity 0.85)
- Time-Evolving Dynamical System for Learning Latent Representations of Mouse Visual Neural Activity (NeurIPS 2025 · similarity 0.85)
- Neural Representational Consistency Emerges from Probabilistic Neural-Behavioral Representation Alignment (ICML 2025 · similarity 0.85)
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.
