Evidence Piece
Evidence piece research focuses on improving the use of evidence in various AI applications, primarily aiming to enhance accuracy, trustworthiness, and explainability. Current research emphasizes developing methods for efficient evidence retrieval and selection, often employing techniques like contrastive learning, language model fine-tuning, and evidence theory, to improve fact-checking, clinical decision support, and other tasks. This work is significant because it addresses critical challenges in AI reliability and transparency, paving the way for more robust and trustworthy AI systems across diverse fields.
Papers
GenAudit: Fixing Factual Errors in Language Model Outputs with Evidence
Kundan Krishna, Sanjana Ramprasad, Prakhar Gupta, Byron C. Wallace, Zachary C. Lipton, Jeffrey P. Bigham
BIDER: Bridging Knowledge Inconsistency for Efficient Retrieval-Augmented LLMs via Key Supporting Evidence
Jiajie Jin, Yutao Zhu, Yujia Zhou, Zhicheng Dou
What Evidence Do Language Models Find Convincing?
Alexander Wan, Eric Wallace, Dan Klein