Reducing Hallucination in Enterprise AI Workflows via Hybrid Utility Minimum Bayes Risk (HUMBR)
Authors:
- Chenhao FangCFChenhao Fang
- Jordi MolaJMJordi Mola
- Mark HarmanMHMark Harman
- Jason NawrockiJNJason Nawrocki
- Vaibhav ShrivastavaVSVaibhav Shrivastava
- Yue ChengYCYue Cheng
- Jay ShahJSJay Shah
- Katayoun ZandKZKatayoun Zand
- Mansi TripathiMTMansi Tripathi
- Arya PudotaAPArya Pudota
- Matthew BeckerMBMatthew Becker
- Hervé RobertHRHervé Robert
Open Access
- 8 August 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 7256-7266
Abstract
No abstract availableAll Related Versions
This publication has 8 references indexed in Scilit:
- LLM-Ensemble: Optimal Large Language Model Ensemble Method for E-commerce Product Attribute Value Extraction2024
- Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate2024
- Chain-of-Verification Reduces Hallucination in Large Language Models2024
- Survey of Hallucination in Natural Language GenerationACM Computing Surveys, 2023
- Epsilon Sampling Rocks: Investigating Sampling Strategies for Minimum Bayes Risk Decoding for Machine Translation2023
- Follow the Wisdom of the Crowd: Effective Text Generation via Minimum Bayes Risk Decoding2023
- Training Language Models to Follow Instructions with Human Feedback2022
- Minimum Bayes-Risk Methods in Automatic Speech Recognition2003