Hallucinations in text generation occur when the system produces text that is not grounded in the input. In this work, we tackle the problem of hallucinations in neural chart summarization. Our analysis shows that the target side of chart summarization training datasets often contains additional information, leading to hallucinations. We propose a natural language inference (NLI) based method to preprocess the training data and show through human evaluation that our method significantly reduces hallucinations. We also found that shortening long-distance dependencies in the input sequence and adding chart-related information like title and legends improves the overall performance.
ChartThinker: A Contextual Chain-of-Thought Approach to Optimized Chart Summarization
Sun Yat-sen University(中山大学);Alibaba Group(阿里巴巴集团);Guangdong Provincial Key Laboratory of Fire Science and Intelligent Emergency Technology(广东省消防科学与智能应急技术重点实验室);Pazhou Lab(琶洲实验室)
Dual-stage and Lightweight Patient Chart Summarization for Emergency Physicians
Department of Electrical and Software Engineering, University of Calgary, Calgary, AB, Canada(电气与软件工程系,卡尔加里大学);Department of Emergency Medicine, University of Calgary, Calgary, AB, Canada(急诊医学系,卡尔加里大学);School of Information, University of Michigan, Ann Arbor, MI, USA(信息学院,密歇根大学);Rockview General Hospital, Calgary, AB, Canada(罗克维尔总医院)