Jay Pujara

Yifan Jiang, Meige Yang, Zitong Li et al.

Affordance reasoning, the inference of an object's action possibilities from its physical properties (e.g., shape and material), is fundamental to human physical understanding and increasingly critical for Large Language Models (LLMs). However, existing affordance benchmarks largely expose explicit object identities in the evaluation setup, allowing models to rely on memorized object-affordance mappings rather than reasoning over physical properties. To address this gap, we introduce Affordance20Q, a novel affordance reasoning benchmark formulated as a 20-Questions game without exposing the object's identity. In each game, the model identifies a hidden object's affordance from a candidate set by asking yes/no questions about its physical properties. Affordance20Q comprises 1,009 games over 454 objects and 59 affordances, all manually filtered, refined, and annotated. We conduct comprehensive experiments with 15 state-of-the-art LLMs and find a substantial gap (~20 points) compared to human performance. A KL-based information-gain (IG) analysis further shows that models fail to ask discriminating questions as the game progresses. To close the gap, we develop KB-Anchored Rule Induction (KARI), a pipeline based on LLMs that generates affordance rules grounded in evidence from knowledge bases (KBs). KARI improves open-source LLMs by up to 15.2 points, while the limited coverage of KBs hinders further gains. We release all our code and data at https://github.com/1171-jpg/Affordance20Q.git

Casandra Rusti, Hussain Hussain, Kian Ahrabian et al.

Institutional prestige shapes access to resources, visibility, and collaboration opportunities in science. Yet whether prestige benefits researchers equally, and how it relates to differences in scientific productivity and collaboration, remains unclear. Here, we quantify prestige advantage as the relative likelihood that researchers at higher-ranked institutions have more collaborators and produce more high-impact papers compared to their lower-ranked peers. Analyzing nearly 5 million papers by 6.5 million authors across more than 65,000 institutions, we present a distributional, tail-sensitive framework to compare prestige advantage across groups. We find that the association between prestige and scientific achievement differs systematically by gender. While both men and women benefit from prestige, the returns are not gender-neutral: women experience comparable advantages only at the most elite institutions, whereas men retain persistent advantages across the broader hierarchy, with disparities widening at higher levels of achievement. Prestige advantage also grows nonlinearly, disproportionately benefiting authors at the most elite institutions. These differences align with collaboration patterns: women's networks are more locally clustered and focused on their own institution, while men collaborate more broadly across institutional strata. Together, these findings reveal a tilted playing field in science: one where prestige amplifies success unevenly and network structure shapes who can access its benefits.