Research map / Time series / Causal and generative time series
On Contrastive Representations of Stochastic Processes
Emile Mathieu, Adam Foster, Yee Whye Teh
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 targeted CReSP instead of CNP | Snooker overlap prediction · Snooker | accuracy (%): 85.3 -> 96.8 | Improved |
| Use targeted CReSP instead of FCLR | Snooker overlap prediction · Snooker | accuracy (%): 85.6 -> 96.8 | Improved |
| Use targeted CReSP instead of MetaCDE | Snooker overlap prediction · Snooker | accuracy (%): 87.7 -> 96.8 | Improved |
| Add self-attention to context aggregation (ablation) | CReSP on sinusoidal parameter recovery · Sinusoids | MSE | Improved |
| Add self-attention to context aggregation (ablation) | CReSP versus FCLR · ShapeNet | test accuracy | Improved |
| Increase training views from 6 to 12 (ablation) | CReSP · ShapeNet | test accuracy | Hurt |
| Use untargeted rather than targeted CReSP representations (ablation) | CReSP · ShapeNet object classification | test accuracy | Improved |
| Use contrastive CReSP instead of reconstructive CNP or ACNP | ShapeNet object classification · ShapeNet | test accuracy | Improved |
| Increase colour distortion (ablation) | CNP and ACNP · ShapeNet object classification | test accuracy | Hurt |
| Apply mild colour distortion (ablation) | CReSP · ShapeNet object classification | test accuracy | Improved |
| Pre-train with CReSP before fitting the downstream classifier | ShapeNet object classification · ShapeNet | test accuracy | Improved |
| Increase the number of test views (ablation) | CReSP and comparison methods · ShapeNet object classification | test accuracy | Improved |
Closest papers by meaning
- Inference via Interpolation: Contrastive Representations Provably Enable Planning and Inference (NeurIPS 2024 · similarity 0.87)
- Learning invariant representations of time-homogeneous stochastic dynamical systems (ICLR 2024 · similarity 0.86)
- Latent Representation and Simulation of Markov Processes via Time-Lagged Information Bottleneck (ICLR 2024 · similarity 0.84)
- Stochastic Process Learning via Operator Flow Matching (NeurIPS 2025 · similarity 0.84)
- Extracting Governing Equations from Latent Dynamics via Multi-View Contrastive Learning (NeurIPS 2026 · similarity 0.84)
- Contrast encodes inductive bias: separating slow noise from dynamics in predictive representation learning (NeurIPS 2026 · 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.
