EMNLP paper: language diffusion models are associative memories with a sharp memorization-generalization transition

LucaAmb · x · 2026-09-08

An IBM/RPI collaboration accepted to EMNLP 2026 Main shows that Uniform-based Discrete Diffusion Models (UDDMs) fundamentally behave as associative memories that store data via basins of attraction—formed by conditional likelihood maximization rather than an explicit energy function.

The paper quantitatively answers when language diffusion models memorize training data and offers a practical metric for assessing memorization risk.

Related event: EMNLP Paper Shows Language Diffusion Models Are Associative Memories(2 posts)→

Original post →

More from Research

Research channel →