CoChat
Sign Up
Cited by 27Topic ModelingInformation Retrieval and Search BehaviorSpeech and dialogue systems

RetGen: A Joint Framework for Retrieval and Grounded Text Generation Modeling

Open access
Zhang, Yizhe, Sun, Siqi, Gao, Xiang, Fang, Yuwei, Brockett, Chris, Galley, Michel, Gao, Jianfeng, Dolan, Bill·2022·Proceedings of the AAAI Conference on Artificial Intelligence
DOI 10.1609/aaai.v36i10.21429
Open this paper in CoChat — free account

Abstract

Recent advances in large-scale pre-training such as GPT-3 allow seemingly high quality text to be generated from a given prompt. However, such generation systems often suffer from problems of hallucinated facts, and are not inherently designed to incorporate useful external information. Grounded generation models appear to offer remedies, but their training typically relies on rarely-available parallel data where information-relevant documents are provided for context. We propose a framework that alleviates this data constraint by jointly training a grounded generator and document retriever on the language model signal. The model learns to reward retrieval of the documents with the highest utility in generation, and attentively combines them using a Mixture-of-Experts (MoE) ensemble to generate follow-on text. We demonstrate that both generator and retriever can take advantage of this joint training and work synergistically to produce more informative and relevant text in both prose and dialogue generation.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1609/aaai.v36i10.21429

Cites

6 of 6 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Recent studies have started exploring joint modeling of retrieval and generation in knowledge-intensive tasks, such as question answering and dialogue generation.

    “Recent studies have started exploring the joint modeling of retrieval and generation in knowledgeintensive tasks, such as question answering (Guu et al., 2020;Lewis et al., 2020;Izacard et al., 2022) and dialogue generation (Zhang et al., 2022).”

    From RepoCoder: Repository-Level Code Completion Through Iterative Retrieval and Generation (2023) · cites this paper as Zhang et al., 2022

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 42

  1. Brown, T. B.; Mann, B.; Ryder, N.; Subbiah, M.; Kaplan, J.; and et al., P. D. 2020. Language Models are Few-Shot Learners. arXiv .
  2. Cai, D.; Wang, Y.; Bi, W.; Tu, Z.; Liu, X.; Lam, W.; and Shi, S. 2019a. Skeleton-to-Response: Dialogue Generation Guided by Retrieval Memory. In NAACL , 1219-1228.
  3. Cai, D.; Wang, Y.; Bi, W.; Tu, Z.; Liu, X.; and Shi, S. 2019b. Retrieval-guided Dialogue Response Generation via a Matching-to-Generation Framework. In EMNLP .
  4. Cho, W. S.; Zhang, Y.; Rao, S.; Celikyilmaz, A.; Xiong, C.; Gao, J.; Wang, M.; and Dolan, B. 2020. Contrastive Multidocument Question Generation. In EACL .
  5. Clement, C. B.; Bierbaum, M.; O'Keeffe, K. P.; and Alemi, A. A. 2019. On the Use of ArXiv as a Dataset. arXiv .
  6. Datar, M.; Immorlica, N.; Indyk, P.; and Mirrokni, V. S. 2004. Locality-sensitive hashing scheme based on p-stable distributions. In Proceedings of the twentieth annual symposium on Computational geometry , 253-262.
  7. Devlin, J.; Chang, M.-W.; Lee, K.; and Toutanova, K. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In NAACL .
  8. Dinan, E.; Roller, S.; Shuster, K.; Fan, A.; Auli, M.; and Weston, J. 2019. Wizard of Wikipedia: Knowledge-Powered Conversational agents. In ICLR .
  9. Eyal, M.; Baumel, T.; and Elhadad, M. 2019. Question Answering as an Automatic Evaluation Metric for News Article Summarization. In NAACL . Minneapolis, Minnesota. Ferragina, P.; and Scaiella, U. 2010. TAGME: On-the-Fly Annotation of Short Text Fragments (by Wikipedia Entities). In Proceedings of the 19th ACM International Conference on Information and Knowledge Management , CIKM '10, 1625-1628. New York, NY, USA: Association for Computing Machinery.
  10. Galley, M.; Brockett, C.; Gao, X.; Gao, J.; and Dolan, B. 2019. Grounded response generation task at DSTC7. In AAAI Dialog System Technology Challenges Workshop .
  11. Gao, J.; Peng, B.; Li, C.; Li, J.; Shayandeh, S.; Liden, L.; and Shum, H.-Y. 2020. Robust conversational AI with grounded text generation. arXiv .
  12. Ghazvininejad, M.; Brockett, C.; Chang, M.-W.; Dolan, B.; Gao, J.; tau Yih, W.; and Galley, M. 2018. A KnowledgeGrounded Neural Conversation Model. In AAAI .
  13. Guu, K.; Lee, K.; Tung, Z.; Pasupat, P.; and Chang, M.-W. 2020. REALM: Retrieval-augmented language model pretraining. arXiv .
  14. Hashimoto, T. B.; Guu, K.; Oren, Y.; and Liang, P. 2018. A Retrieve-and-Edit Framework for Predicting Structured Outputs. In NeurIPS .
  15. Huang, L.; Wu, L.; and Wang, L. 2020. Knowledge GraphAugmented Abstractive Summarization with SemanticDriven Cloze Reward. In ACL . Online.
  16. Karpukhin, V.; O˘ guz, B.; Min, S.; Wu, L.; Edunov, S.; Chen, D.; and Yih, W.-t. 2020. Dense Passage Retrieval for OpenDomain Question Answering. arXiv .
  17. Lavie, A.; and Agarwal, A. 2007. METEOR: An automatic metric for MT evaluation with high levels of correlation with human judgments. In Proceedings of the Second Workshop on Statistical Machine Translation , 228-231. Association for Computational Linguistics.
  18. Lewis, M.; Ghazvininejad, M.; Ghosh, G.; Aghajanyan, A.; Wang, S.; and Zettlemoyer, L. 2020a. Pre-training via paraphrasing. NeurIPS .
  19. Lewis, P.; Perez, E.; Piktus, A.; Petroni, F.; Karpukhin, V.; Goyal, N.; K¨ uttler, H.; Lewis, M.; Yih, W.-t.; Rockt¨ aschel, T.; et al. 2020b. Retrieval-augmented generation for knowledgeintensive NLP tasks. arXiv .
  20. Li, J.; Galley, M.; Brockett, C.; Gao, J.; and Dolan, B. 2016. A diversity-promoting objective function for neural conversation models. NAACL .
  21. Li, J.; Jia, R.; He, H.; and Liang, P. 2018. Delete, Retrieve, Generate: A Simple Approach to Sentiment and Style Transfer. In NAACL .
  22. Liu, S.; Chen, H.; Ren, Z.; Feng, Y.; Liu, Q.; and Yin, D. 2018. Knowledge Diffusion for Neural Dialogue Generation. In ACL .
  23. Liu, Z.; Niu, Z.-Y.; Wu, H.; and Wang, H. 2019. Knowledge Aware Conversation Generation with Explainable Reasoning over Augmented Graphs. In EMNLP .
  24. Nelson, B. L. 1990. Control variate remedies. Operations Research , 38(6): 974-992.
  25. Papineni, K.; Roukos, S.; Ward, T.; and Zhu, W.-J. 2002. BLEU: a method for automatic evaluation of machine translation. ACL .
  26. Peng, B.; Li, C.; Li, J.; Shayandeh, S.; Liden, L.; and Gao, J. 2020. SOLOIST: Few-shot Task-Oriented Dialog with A Single Pre-trained Auto-regressive Model. arXiv .
  27. Peng, H.; Parikh, A. P.; Faruqui, M.; Dhingra, B.; and Das, D. 2019. Text Generation with Exemplar-based Adaptive Decoding. In NAACL .
  28. Shrivastava, A.; and Li, P. 2014. Asymmetric LSH (ALSH) for sublinear time maximum inner product search (MIPS). NeurIPS .
  29. Shuster, K.; Poff, S.; Chen, M.; Kiela, D.; and Weston, J. 2021. Retrieval Augmentation Reduces Hallucination in Conversation. arXiv .
  30. Song, Y.; Li, C.-T.; Nie, J.-Y .; Zhang, M.; Zhao, D.; and Yan, R. 2018. An ensemble of retrieval-based and generationbased human-computer conversation systems. In IJCAI , 4382-4388.
  31. Thulke, D.; Daheim, N.; Dugast, C.; and Ney, H. 2021. Efficient Retrieval Augmented Generation from Unstructured Knowledge for Task-Oriented Dialog. arXiv .
  32. Williams, R. J. 1992. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine learning , 8(3-4): 229-256.
  33. Wu, Y.; Wei, F.; Huang, S.; Wang, Y.; Li, Z.; and Zhou, M. 2019. Response generation by context-aware prototype editing. In AAAI , 7281-7288.
  34. Wu, Z.; Galley, M.; Brockett, C.; Zhang, Y.; Gao, X.; Quirk, C.; Koncel-Kedziorski, R.; Gao, J.; Hajishirzi, H.; Ostendorf, M.; and Dolan, B. 2021. A Controllable Model of Grounded Response Generation. In AAAI .
  35. Xiong, L.; Xiong, C.; Li, Y.; Tang, K.-F.; Liu, J.; Bennett, P.; Ahmed, J.; and Overwijk, A. 2020. Approximate nearest neighbor negative contrastive learning for dense text retrieval. arXiv .
  36. Yang, L.; Hu, J.; Qiu, M.; Qu, C.; Gao, J.; Croft, W. B.; Liu, X.; Shen, Y.; and Liu, J. 2019. A Hybrid Retrieval-Generation Neural Conversation Model. In Proceedings of the 28th ACM International Conference on Information and Knowledge Management , 1341-1350. ACM.
  37. Young, T.; Cambria, E.; Chaturvedi, I.; Huang, M.; Zhou, H.; and Biswas, S. 2018. Augmenting end-to-end dialogue systems with commonsense knowledge. In AAAI .
  38. Zhang, Y.; Galley, M.; Gao, J.; Gan, Z.; Li, X.; Brockett, C.; and Dolan, B. 2018. Generating informative and diverse conversational responses via adversarial information maximization. NeurIPS .
  39. Zhang, Y.; Sun, S.; Galley, M.; Chen, Y.-C.; Brockett, C.; Gao, X.; Gao, J.; Liu, J.; and Dolan, B. 2020. DialoGPT: LargeScale Generative Pre-training for Conversational Response Generation. In ACL, system demonstration .
  40. Zhao, X.; Wu, W.; Xu, C.; Tao, C.; Zhao, D.; and Yan, R. 2020. Knowledge-grounded dialogue generation with pretrained language models. arXiv preprint arXiv:2010.08824 . · arXiv
  41. Zhou, K.; Prabhumoye, S.; and Black, A. W. 2018. A Dataset for Document Grounded Conversations. In EMNLP , 708713.
  42. Zhu, W.; Mo, K.; Zhang, Y.; Zhu, Z.; Peng, X.; and Yang, Q. 2017. Flexible End-to-End Dialogue System for Knowledge Grounded Conversation. arXiv .

Cited by

Sorted by citations

Authors
Zhang, Y., Sun, S., Gao, X., Fang, Y., Brockett, C., Galley, M., Gao, J., & Dolan, B.
Affiliations
Microsoft (United States) (US)
ORCID
Venue
Proceedings of the AAAI Conference on Artificial Intelligence
Publisher
Association for the Advancement of Artificial Intelligence (AAAI)
Published
2022-06-28
Volume
36
Issue
10
Pages
11739-11747
ISSN
  • 2159-5399 (ISSN-L)
  • 2374-3468
DOI
10.1609/aaai.v36i10.21429
OpenAlex
W4206890427
Cited by
27
References
6
Citations by year
2026: 2 · 2025: 7 · 2024: 4 · 2023: 11 · 2022: 3
FWCI
3.45
Type
conference paper
Language
en
Open access
Locations
  • Proceedings of the AAAI Conference on Artificial Intelligence (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)) (submittedVersion)
  • arXiv (Cornell University)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
  • Physical Sciences › Computer Science › Artificial Intelligence › Speech and dialogue systems
Keywords
retrieval-augmented generation, document retrieval, hallucination mitigation, mixture of experts, dialogue generation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review