WRSWorld Research StateContribute
WORLD RESEARCH STATE

Hinton: training products of experts by minimizing contrastive divergence

Contribution grade

Generative and foundation models · ALGORITHM

Reference score175,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Introduced contrastive divergence as a tractable approximate learning objective and demonstrated product-of-experts training.

Scientific scope

The record is Hinton's sole-author 2002 paper and its product-of-experts experiments. Contrastive divergence is an approximation, not exact maximum-likelihood training or a convergence guarantee for every energy-based model.

Evidence

The linked sources identify this contribution and its scope. WRS has not independently reconstructed the scientific claims. The record is Hinton's sole-author 2002 paper and its product-of-experts experiments. Contrastive divergence is an approximation, not exact maximum-likelihood training or a convergence guarantee for every energy-based model.

Contributors

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction