Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
60 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
8 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Seeing is Believing: Mitigating Hallucination in Large Vision-Language Models via CLIP-Guided Decoding (2402.15300v2)

Published 23 Feb 2024 in cs.CV, cs.AI, cs.CL, cs.LG, and cs.MM

Abstract: Large Vision-LLMs (LVLMs) are susceptible to object hallucinations, an issue in which their generated text contains non-existent objects, greatly limiting their reliability and practicality. Current approaches often rely on the model's token likelihoods or other internal information, instruction tuning on additional datasets, or incorporating complex external tools. We first perform empirical analysis on sentence-level LVLM hallucination, finding that CLIP similarity to the image acts as a stronger and more robust indicator of hallucination compared to token likelihoods. Motivated by this, we introduce our CLIP-Guided Decoding (CGD) approach, a straightforward but effective training-free approach to reduce object hallucination at decoding time. CGD uses CLIP to guide the model's decoding process by enhancing visual grounding of generated text with the image. Experiments demonstrate that CGD effectively mitigates object hallucination across multiple LVLM families while preserving the utility of text generation. Codes are available at https://github.com/d-ailin/CLIP-Guided-Decoding.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (57)
  1. Gpt-4 technical report. arXiv preprint arXiv:2303.08774.
  2. Towards a human-like open-domain chatbot. arXiv preprint arXiv:2001.09977.
  3. nocaps: novel object captioning at scale. In Proceedings of the IEEE International Conference on Computer Vision, pages 8948–8957.
  4. Spice: Semantic propositional image caption evaluation. Computer Vision–ECCV 2016, pages 382–398.
  5. Why exposure bias matters: An imitation learning perspective of error accumulation in language generation. arXiv preprint arXiv:2204.01171.
  6. METEOR: An automatic metric for MT evaluation with improved correlation with human judgments. In Goldstein, J., Lavie, A., Lin, C.-Y., and Voss, C., editors, Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization, pages 65–72, Ann Arbor, Michigan. Association for Computational Linguistics.
  7. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3).
  8. Rt-2: Vision-language-action models transfer web knowledge to robotic control. arXiv preprint arXiv:2307.15818.
  9. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality. See https://vicuna. lmsys. org (accessed 14 April 2023).
  10. Dola: Decoding by contrasting layers improves factuality in large language models. arXiv preprint arXiv:2309.03883.
  11. Instructblip: Towards general-purpose vision-language models with instruction tuning. arXiv preprint arXiv:2305.06500.
  12. Data filtering networks. arXiv preprint arXiv:2309.17425.
  13. Eva: Exploring the limits of masked visual representation learning at scale. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19358–19369.
  14. Detecting and preventing hallucinations in large vision language models. arXiv preprint arXiv:2308.06394.
  15. Clipscore: A reference-free evaluation metric for image captioning. arXiv preprint arXiv:2104.08718.
  16. Sugarcrepe: Fixing hackable benchmarks for vision-language compositionality. In Thirty-Seventh Conference on Neural Information Processing Systems Datasets and Benchmarks Track.
  17. Opera: Alleviating hallucination in multi-modal large language models via over-trust penalty and retrospection-allocation. arXiv preprint arXiv:2311.17911.
  18. Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12):1–38.
  19. Language models (mostly) know what they know. arXiv preprint arXiv:2207.05221.
  20. Deep visual-semantic alignments for generating image descriptions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3128–3137.
  21. Openimages: A public dataset for large-scale multi-label and multi-class image classification. Dataset available from https://github. com/openimages, 2(3):18.
  22. Mitigating object hallucinations in large vision-language models through visual contrastive decoding. arXiv preprint arXiv:2311.16922.
  23. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. arXiv preprint arXiv:2301.12597.
  24. Evaluating object hallucination in large vision-language models. In Bouamor, H., Pino, J., and Bali, K., editors, Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 292–305, Singapore. Association for Computational Linguistics.
  25. Lin, C.-Y. (2004). ROUGE: A package for automatic evaluation of summaries. In Text Summarization Branches Out, pages 74–81, Barcelona, Spain. Association for Computational Linguistics.
  26. Microsoft coco: Common objects in context. In Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 13, pages 740–755. Springer.
  27. Mitigating hallucination in large multi-modal models via robust instruction tuning. arXiv preprint arXiv:2306.14565, 1(2):9.
  28. Visual instruction tuning.
  29. Selfcheckgpt: Zero-resource black-box hallucination detection for generative large language models. arXiv preprint arXiv:2303.08896.
  30. Bleu: a method for automatic evaluation of machine translation. In Isabelle, P., Charniak, E., and Lin, D., editors, Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311–318, Philadelphia, Pennsylvania, USA. Association for Computational Linguistics.
  31. Learning transferable visual models from natural language supervision. In International conference on machine learning, pages 8748–8763. PMLR.
  32. Sequence level training with recurrent neural networks. arXiv preprint arXiv:1511.06732.
  33. Object hallucination in image captioning. In Riloff, E., Chiang, D., Hockenmaier, J., and Tsujii, J., editors, Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 4035–4045, Brussels, Belgium. Association for Computational Linguistics.
  34. Xraygpt: Chest radiographs summarization using medical vision-language models. arXiv preprint arXiv:2306.07971.
  35. Winoground: Probing vision and language models for visio-linguistic compositionality. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5238–5248.
  36. Eyes wide shut? exploring the visual shortcomings of multimodal llms. arXiv preprint arXiv:2401.06209.
  37. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971.
  38. Cider: Consensus-based image description evaluation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4566–4575.
  39. Vigc: Visual instruction generation and correction. arXiv preprint arXiv:2308.12714.
  40. Sam-clip: Merging vision foundation models towards semantic and spatial understanding. arXiv preprint arXiv:2310.15308.
  41. Evaluation and analysis of hallucination in large vision-language models. arXiv preprint arXiv:2308.15126.
  42. Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2097–2106.
  43. Google’s neural machine translation system: Bridging the gap between human and machine translation. arXiv preprint arXiv:1609.08144.
  44. The rise and potential of large language model based agents: A survey. arXiv preprint arXiv:2309.07864.
  45. Can llms express their uncertainty? an empirical evaluation of confidence elicitation in llms. arXiv preprint arXiv:2306.13063.
  46. Demystifying clip data. arXiv preprint arXiv:2309.16671.
  47. mplug-owl: Modularization empowers large language models with multimodality. arXiv preprint arXiv:2304.14178.
  48. Woodpecker: Hallucination correction for multimodal large language models. arXiv preprint arXiv:2310.16045.
  49. Mm-vet: Evaluating large multimodal models for integrated capabilities. arXiv preprint arXiv:2308.02490.
  50. Mmmu: A massive multi-discipline multimodal understanding and reasoning benchmark for expert agi. arXiv preprint arXiv:2311.16502.
  51. When and why vision-language models behave like bags-of-words, and what to do about it? In The Eleventh International Conference on Learning Representations.
  52. From recognition to cognition: Visual commonsense reasoning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6720–6731.
  53. Sigmoid loss for language image pre-training. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 11975–11986.
  54. How language model hallucinations can snowball. arXiv preprint arXiv:2305.13534.
  55. Calibrating sequence likelihood improves conditional language generation. In The Eleventh International Conference on Learning Representations.
  56. Unified vision-language pre-training for image captioning and vqa. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 13041–13049.
  57. Analyzing and mitigating object hallucination in large vision-language models. arXiv preprint arXiv:2310.00754.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Ailin Deng (11 papers)
  2. Zhirui Chen (11 papers)
  3. Bryan Hooi (158 papers)