Feature Attribution Benefits in Supervised Contrastive Learning
Explore how supervised contrastive learning improves feature attribution for better model interpretability, robustness, and trustworthiness in AI systems.
FeatEHR-LLM: AI-Driven Feature Engineering for EHR Data
Discover FeatEHR-LLM, a novel AI framework using Large Language Models to improve feature engineering in Electronic Health Records for better clinical pred...
CGC: Enhancing Fine-Grained Multi-Image Understanding
Discover how CGC improves fine-grained multi-image understanding in MLLMs using contrastive learning and spatial rewards for better AI accuracy.
SSG: Enhanced Logit-Balanced Watermarking for LLMs
Discover SSG, a logit-balanced vocabulary partitioning method that boosts watermark strength and detectability in large language models.
Deciding Fact Relevance in Boolean Conjunctive Queries
Explore the complexity of determining fact relevance in Boolean conjunctive queries and how structural constraints simplify this challenge.