October 30, 2025 front

ML Newsnew | past | best | rss | submit

Stories from October 30, 2025 (UTC)

Go back a day, month, or year. Go forward a day.

1.	Language models are injective and hence invertible (arxiv.org)
	231 points by mazsa 339 days ago \| hide \| past \| pdf \| 147 comments
2.	Multi-Domain Rubrics Requiring Professional Knowledge to Answer and Judge (arxiv.org)
	2 points by PaulHoule 338 days ago \| hide \| past \| pdf \| discuss
3.	Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity (arxiv.org)
	2 points by CharlesW 338 days ago \| hide \| past \| pdf \| discuss
4.	Predictability Enables Parallelization (arxiv.org)
	2 points by leokoz8 339 days ago \| hide \| past \| pdf \| discuss
5.	Drug Discovery with Apex Protocol by Nvidia and Numerion Labs (arxiv.org)
	2 points by nkko 339 days ago \| hide \| past \| pdf \| discuss
6.	Benchmarking World-Model Learning (arxiv.org)
	2 points by mpweiher 339 days ago \| hide \| past \| pdf \| discuss