Research map / Probabilistic and scientific modeling / Variational and generative sampling
Conditional Noise-Contrastive Estimation of Unnormalised Models
Read the paperCheck 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 |
|---|---|---|---|
| Generate contrastive noise conditionally from each data point rather than independently | CNCE vs NCE on the ring model · Ring model | estimation error: about one order of magnitude lower | Helped |
| Generate contrastive noise conditionally from each data point rather than independently | CNCE vs NCE on the Gaussian model · Gaussian model | estimation error | Hurt |
| Generate contrastive noise conditionally from each data point rather than independently | CNCE vs NCE on the ICA model · ICA model | estimation error | No clear effect |
| Increase the number of noise samples per data point | CNCE and NCE · Gaussian and ICA models | estimation error difference between methods | No clear effect |
| Increase the number of data points | CNCE on Gaussian and ICA models · Gaussian and ICA models | estimation error | Helped |
| Use a small conditional-noise scale so noise samples are close to data points | CNCE · Unnormalised density estimation | score-matching objective correspondence: ε2/2 | Helped |
Closest papers by meaning
- A Unified View on Learning Unnormalized Distributions via Noise-Contrastive Estimation (ICML 2025 · similarity 0.89)
- Learning normalized image densities via dual score matching (NeurIPS 2025 · similarity 0.85)
- Label-Noise Robust Diffusion Models (ICLR 2024 · similarity 0.85)
- Zero-Shot Conditioning of Score-Based Diffusion Models by Neuro-Symbolic Constraints (AAAI 2025 · similarity 0.84)
- Sampling Binary Data by Denoising through Score Functions (ICML 2025 · similarity 0.84)
- Robust Representation Consistency Model via Contrastive Denoising (ICLR 2025 · similarity 0.84)
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.