Research map / Probabilistic and scientific modeling / Dynamical systems
CHESS: Chebyshev Spectral Synthesis for Trajectory Condensation
Ruituo Wu, Hongyu Zhang, Qiang Wang, Jiawei Du, Wei Cui, Ce Zhu, Bing Li
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 |
|---|---|---|---|
| add piecewise Chebyshev fitting (ablation) | CHESS · MeR (SPC=5) MobileNet | accuracy: 62.2 -> 71.6 | Improved |
| add piecewise Chebyshev fitting (ablation) | CHESS · UCI-HAR (SPC=1) ShuffleNet | accuracy: 52.2 -> 61.6 | Improved |
| add low-rank SVD constraint (ablation) | CHESS · MeR ResNet / UCI-HAR THAT | accuracy: 72.5 -> 73.2 (MeR ResNet); 63.3 -> 62.9 (UCI-HAR THAT) | Mixed |
| use Chebyshev basis instead of monomial Poly basis (ablation) | CHESS · MeR ResNet target | accuracy: 57.1 -> 73.2 | Improved |
| use Chebyshev basis instead of monomial Poly basis (ablation) | CHESS · MeR ConvNet | accuracy: 65.3 -> 77.1 | Improved |
| swap Chebyshev for Legendre or Hermite basis (ablation) | CHESS · MeR ResNet target | accuracy: 73.2 vs 73.2 (Legendre) and 73.1 (Hermite) | No clear effect |
| add least-squares manifold initialization (ablation) | CHESS · MeR ResNet target | accuracy: 14.4 -> 73.2 | Improved |
| use analytic Chebyshev resampling instead of interpolation (ablation) | CHESS vs DANCE · MeR zero-shot length 2000 | accuracy: 62.3 -> 75.8 | Improved |
| use analytic Chebyshev resampling instead of downsampling | CHESS vs DANCE · MeR zero-shot length 200 | accuracy: 66.3 -> 76.9 | Improved |
| function-first synthesis instead of pixel-wise DM | CHESS vs NCFM · ActR (SPC=5) | accuracy: 56.9 -> 71.0 | Improved |
| CHESS compared with DC on Transformer backbone | CHESS vs DC · UCI-HAR (SPC=1) THAT | accuracy: 64.0 -> 62.9 | Hurt |
| increase polynomial degree beyond 2-3 (ablation) | CHESS · all datasets | accuracy | No clear effect |
Closest papers by meaning
- StreamPhy: Streaming Inference of High-Dimensional Physical Dynamics via State Space Models (NeurIPS 2026 · similarity 0.84)
- Christoffel-DPS: Optimal sensor placement in diffusion posterior sampling for arbitrary distributions (NeurIPS 2026 · similarity 0.84)
- Markovian Dynamics Enforcer: Feasibility Preserving Correction on Learned Dyanmics Manifolds (NeurIPS 2026 · similarity 0.83)
- Spectral Alignment in Forward–Backward Representations via Temporal Abstraction (NeurIPS 2026 · similarity 0.83)
- Optimal Sensor Scheduling and Selection for Continuous-Discrete Kalman Filtering with Auxiliary Dynamics (ICML 2025 · similarity 0.83)
- Learning Robust Spectral Dynamics for Temporal Domain Generalization (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.
