Research map / Probabilistic and scientific modeling / Dynamical systems
Extraction and Recovery of Spatio-Temporal Structure in Latent Dynamics Alignment with Diffusion Models
Yule Wang, Zijing Wu, Chengrui Li, Anqi Wu
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 |
|---|---|---|---|
| remove spatial transformer layer from the diffusion model (ablation) | ERDiff · NHP cross-day | R2 (%): 18.81 -> -12.69 | Hurt |
| remove temporal transformer layer from the diffusion model (ablation) | ERDiff · NHP cross-day | R2 (%): 18.81 -> -14.61 | Hurt |
| remove spatial transformer layer (ablation) | ERDiff · NHP inter-subject | R2 (%): 10.29 -> -14.60 | Hurt |
| remove temporal transformer layer (ablation) | ERDiff · NHP inter-subject | R2 (%): 10.29 -> -17.10 | Hurt |
| remove spatial transformer layer (ablation) | ERDiff · NHP cross-day | RMSE: 7.99 -> 8.57 | Hurt |
| remove temporal transformer layer (ablation) | ERDiff · NHP cross-day | RMSE: 7.99 -> 8.93 | Hurt |
| use ERDiff instead of JSDM metric-based alignment | JSDM · NHP cross-day | R2 (%): -17.36 -> 18.81 | Improved |
| use ERDiff instead of DAF adversarial alignment | DAF · NHP cross-day | R2 (%): -6.37 -> 18.81 | Improved |
| use ERDiff instead of DAF | DAF · NHP inter-subject | RMSE: 9.57 -> 9.78 | Mixed |
| use ERDiff instead of RDA-MMD | RDA-MMD · Rat hippocampal CA1 | R2 (%): 21.73 -> 32.69 | Improved |
| use ERDiff instead of Cycle-GAN | Cycle-GAN · NHP cross-day | R2 (%): -24.83 -> 18.81 | Improved |
| reduce target-domain sampling density from 50% to 10% | ERDiff · NHP motor cortex | decoding accuracy | No clear effect |
Closest papers by meaning
- Modeling Latent Neural Dynamics with Gaussian Process Switching Linear Dynamical Systems (NeurIPS 2024 · similarity 0.88)
- Neural Embeddings Rank: Aligning 3D latent dynamics with movements (NeurIPS 2024 · similarity 0.87)
- Latent Diffusion for Neural Spiking Data (NeurIPS 2024 · similarity 0.87)
- Meta-Dynamical State Space Models for Integrative Neural Data Analysis (ICLR 2025 · similarity 0.87)
- Exploring Behavior-Relevant and Disentangled Neural Dynamics with Generative Diffusion Models (NeurIPS 2024 · similarity 0.87)
- Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shifts (NeurIPS 2026 · 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.
