Smoothing the Landscape: Causal Structure Learning via Diffusion Denoising Objectives Work done while at the Department of Computer Science, Tufts University

Hao Zhu, Di Zhou, Donna Slonim · 2026

Read Smoothing the Landscape: Causal Structure Learning via Diffusion Denoising Objectives Work done while at the Department of Computer Science, Tufts University by Hao Zhu, Di Zhou, Donna Slonim (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-02250.txt.

Description

Zhu, Zhou and Slonim smooth causal learning with diffusion denoising goals. Older methods like NOTEARS and DAG-GNN falter on scale and stability when features outnumber samples. The denoising score goal smooths gradients for faster, steadier convergence, and a hop based loop rule runs faster than matrix inversion. Their DDCD frame uses the reverse denoising process to infer causal structure rather than generate data, doing well on synthetic tests and two real examples.

Catalog record

CREATORHao Zhu, Di Zhou, Donna Slonim
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE93911
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.02250

Browse the full library · About the Tower of Babel