3

CausalDS is a benchmark generator for agentic causal data science: every problem is generated fresh in its entirety, with tasks spanning all three rungs of Pearl’s hierarchy that involve significant tool use. Exam composition is a free parameter: it can be tailored to a specific goal or grounded in real-world corpora. The benchmark jointly tests symbolic causal reasoning, data-science execution, uncertainty quantification, epistemic abstention, and coding/tool use.

The classical Tweedie’s formula connects the score under Gaussian noise to the posterior mean. We generalize this result to the Energy-Tweedie identity: the score from a wider family of noising distributions is connected to a path-derivative of a matched energy score. Among other things, this opens new routes to score estimation, noise parameter estimation, and provides the score-based perspective on diffusion approaches based on scoring rules.

Modeling the translation and review processes as zero-inflated fat-tailed distributions, we show how to extract useful information on translators’ reliability with as little as one review per translation.