Hans
PulseAugur coverage of Hans — every cluster mentioning Hans across labs, papers, and developer communities, ranked by signal.
-
New UNMASK pipeline automatically finds and fixes spurious correlations in text classifiers
Researchers have developed UNMASK, an automated pipeline designed to identify and verify spurious correlations in text classifiers. This system discovers potential surface patterns that models exploit without true lingu…
-
New UNMASK system automatically finds and fixes spurious correlations in text classifiers
A new research paper introduces UNMASK, an automated pipeline designed to identify and correct spurious correlations in text classifiers. This system uses causal verification and group-based reweighting to address issue…
-
Scientists cautiously study 'talking dogs' using soundboards to avoid Clever Hans effect
A sheepadoodle named Bunny has learned to use a soundboard with over 100 words, attracting significant online attention and becoming a subject of scientific research. Cognitive scientist Federico Rossano at UC San Diego…
-
Huawei AI Glasses Enhance Accessibility, Integrating Human and AI Support
Huawei is enhancing accessibility through its AI-powered devices, notably the "Xiao Yi Sees the World" AI glasses. These glasses offer real-time environmental descriptions to visually impaired users, assisting with navi…
-
Adaptive negative scheduling framework boosts graph contrastive learning performance
Researchers have introduced AdNGCL, a novel framework for graph contrastive learning designed to overcome the limitations of static negative sampling. This adaptive approach utilizes a hardness-aware scheduler (HANS) to…
-
New framework enhances graph contrastive learning with adaptive negative scheduling
Researchers have introduced AdNGCL, a new framework designed to improve graph contrastive learning (GCL) for self-supervised representation learning. This method addresses the limitations of static negative sampling by …