mlsanityData to 7 Oct 2026

Tackling Hallucinations in Neural Chart Summarization

arXiv 2308.00399 ↗First posted 1 Aug 2023cs.CL cs.LGcited by 6

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 23 of its 36 references that are in the map. First posted on arXiv 1 Aug 2023.

Builds on

23 of 36 references.

References · 23

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  3. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,874
  4. 2022Survey of Hallucination in Natural Language Generationcited by 2,056
  5. 2018A Call for Clarity in Reporting BLEU Scorescited by 1,790
  6. 2020On Faithfulness and Factuality in Abstractive Summarizationcited by 895
  7. 2021Entity-Based Knowledge Conflicts in Question Answeringcited by 349
  8. 2017Six Challenges for Neural Machine Translationcited by 334
  9. 2019Benchmarking Zero-shot Text Classification: Datasets, Evaluation and Entailment Approachcited by 299
  10. 2020ToTTo: A Controlled Table-To-Text Generation Datasetcited by 267
  11. 2022Chart-to-Text: A Large-Scale Benchmark for Chart Summarizationcited by 166
  12. 2021The Curious Case of Hallucinations in Neural Machine Translationcited by 161
11 more
  1. 2021The Factual Inconsistency Problem in Abstractive Text Summarization: A Surveycited by 79
  2. 2020Chart-to-Text: Generating Natural Language Descriptions for Charts by Adapting the Transformer Modelcited by 77
  3. 2021SciCap: Generating Captions for Scientific Figurescited by 76
  4. 2020Evaluating Semantic Accuracy of Data-to-Text Generation with Natural Language Inferencecited by 43
  5. 2021Controlling Hallucinations at Word Level in Data-to-Text Generationcited by 41
  6. 2020NUBIA: NeUral Based Interchangeability Assessor for Text Generationcited by 34
  7. 2020A Gold Standard Methodology for Evaluating Accuracy in Data-To-Text Systemscited by 34
  8. 2020Revisiting Challenges in Data-to-Text Generation with Fact Groundingcited by 27
  9. 2021AutoChart: A Dataset for Chart-to-Text Generation Taskcited by 12
  10. 2021AgreeSum: Agreement-Oriented Multi-Document Summarizationcited by 10
  11. 2021MiRANews: Dataset and Benchmarks for Multi-Resource-Assisted News Summarizationcited by 2

Built on by

Cited by year

Citing papers per year: 2024: 4; 2025: 2.

6 papers.

  1. 2024An Audit on the Perspectives and Challenges of Hallucinations in NLPcited by 17
  2. 2025ChartInsighter: An Approach for Mitigating Hallucination in Time-series Chart Summary Generation with A Benchmark Datasetcited by 9
  3. 2024Natural Language Generation for Visualizations: State of the Art, Challenges and Future Directionscited by 5
  4. 2025How Much Do LLMs Hallucinate across Languages? On Realistic Multilingual Estimation of LLM Hallucinationcited by 5
  5. 2024An Extensive Evaluation of Factual Consistency in Large Language Models for Data-to-Text Generationcited by 2
  6. 2024Impact of Model Size on Fine-tuned LLM Performance in Data-to-Text Generation: A State-of-the-Art Investigationcited by 1