# A Diffusion Theory For Deep Learning Dynamics: Stochastic Gradient Descent Exponentially Favors Flat Minima

> Research article (International Conference on Learning Representations, 2021) · cited 16× · AI/ML

**Wikidata**: [openalex:W3118431319](https://www.wikidata.org/wiki/openalex:W3118431319)  
**Source**: https://4ort.xyz/entity/a-diffusion-theory-for-deep-learning-dynamics-stochastic-gradient-descent-exponentially-favors-flat-minima