CoChat
Sign Up
Cited by 40Domain Adaptation and Few-Shot LearningTopic ModelingAdvanced Neural Network Applications

The DOI matches its Crossref record; 3 of the 3 references it cites were found in a registry or an index.

Aging with GRACE: Lifelong Model Editing with Discrete Key-Value Adaptors

Open access
Hartvigsen, Tom, Sankaranarayanan, Swami, Palangi, Hamid, Kim, Yoon, Ghassemi, Marzyeh·2023·Advances in Neural Information Processing Systems 36
DOI 10.52202/075280-2079
Open this paper in CoChat — free accountOpen PDF

Abstract

Deployed language models decay over time due to shifting inputs, changing user needs, or emergent world-knowledge gaps. When such problems are identified, we want to make targeted edits while avoiding expensive retraining. However, current model editors, which modify such behaviors of pre-trained models, degrade model performance quickly across multiple, sequential edits. We propose GRACE, a lifelong model editing method, which implements spot-fixes on streaming errors of a deployed model, ensuring minimal impact on unrelated inputs. GRACE writes new mappings into a pre-trained model's latent space, creating a discrete, local codebook of edits without altering model weights. This is the first method enabling thousands of sequential edits using only streaming errors. Our experiments on T5, BERT, and GPT models show GRACE's state-of-the-art performance in making and retaining edits, while generalizing to unseen inputs. Our code is available at https://www.github.com/thartvigsen/grace}.

Overview

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 · 10.52202/075280-2079

Cites

0 citation errors · 0 not checkable

Statistics

Verification

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?

  • SERAC/DEFER uses working memory stored in additional small models: a scope classifier and a counterfactual model, whose knowledge is parametric.

    “It is worth noting that SERAC [32]/DEFER [10] uses working memory that is stored in additional small models: a scope classifier and a counterfactual model, whose knowledge is parametric.”

    From WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models (2024) · cites this paper as [10]

    Appendix C (p. 16) describes Defer as using two trained one-layer networks: one predicts whether to defer, and the other predicts a replacement value. This supports the claim's description of additional parametric models for Defer, but the paper calls Defer an altered method inspired by SERAC (Experimental Setup, p. 5); it does not establish that both methods store working memory in a scope classifier and counterfactual model. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • GRACE is designed for lifelong editing via retrieval-based working memory.

    “GRACE [10] is designed for lifelong editing via retrieval-based working memory.”

    From WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models (2024) · cites this paper as [10]

  • In-context learning and retrieval-based editing methods like GRACE contribute to working memory.

    “In-context learning and retrieval-based editing methods like GRACE contribute to working memory [60, 10].”

    From WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models (2024) · cites this paper as [10]

  • Current knowledge injection methods often suffer from computational overhead, catastrophic forgetting, and overfitting.

    “current knowledge injection methods often suffer from computational overhead [13, 10], catastrophic forgetting [63], and overfitting [64].”

    From WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models (2024) · cites this paper as [10]

  • Retrieval-based methods like SERAC and GRACE improve working memory for editing.

    “Retrieval-based methods like SERAC and GRACE improve working memory for editing [32, 10].”

    From WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Models (2024) · cites this paper as [10]

Self-check

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

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

Citations

Cites

The publisher hasn’t deposited this paper’s reference list.

Research works printed but not linked: 45

  1. [1][1] Rahaf Aljundi, Lucas Caccia, Eugene Belilovsky, Massimo Caccia, Min Lin, Laurent Charlin, and Tinne Tuytelaars. Online continual learning with maximally interfered retrieval. ArXiv preprint, abs/1908.04742 , 2019. - · arXiv
  2. [3][3] Stella Biderman, Hailey Schoelkopf, Quentin Anthony, Herbie Bradley, Kyle O'Brien, Eric Hallahan, Mohammad Aflah Khan, Shivanshu Purohit, USVSN Sai Prashanth, Edward Raff, et al. Pythia: A suite for analyzing large language models across training and scaling. arXiv preprint arXiv:2304.01373 , 2023. - · arXiv
  3. [4][4] Davis Brown, Charles Godfrey, Cody Nizinski, Jonathan Tu, and Henry Kvinge. Robustness of edited neural networks. arXiv preprint arXiv:2303.00046 , 2023. - · arXiv
  4. [5][5] Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in neural information processing systems , 33:1877-1901, 2020. -
  5. [6][6] Ilias Chalkidis, Tommaso Pasini, Sheng Zhang, Letizia Tomada, Sebastian Felix Schwemer, and Anders Søgaard. Fairlex: A multilingual benchmark for evaluating fairness in legal text processing. arXiv preprint arXiv:2203.07228 , 2022. - · arXiv
  6. [7][7] Yi Dai, Hao Lang, Yinhe Zheng, Fei Huang, Luo Si, and Yongbin Li. Lifelong learning for question answering with hierarchical prompts. arXiv preprint arXiv:2208.14602 , 2022. - · arXiv
  7. [8][8] Nicola De Cao, Wilker Aziz, and Ivan Titov. Editing factual knowledge in language models. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing , pages 6491-6506, 2021. -
  8. [9][9] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 , 2020. - · arXiv
  9. [11][11] Thomas Hartvigsen, Saadia Gabriel, Hamid Palangi, Maarten Sap, Dipankar Ray, and Ece Kamar. Toxigen: A large-scale machine-generated dataset for adversarial and implicit hate speech detection. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages 3309-3326, 2022. -
  10. [12][12] Peter Hase, Mona Diab, Asli Celikyilmaz, Xian Li, Zornitsa Kozareva, Veselin Stoyanov, Mohit Bansal, and Srinivasan Iyer. Do language models have beliefs? methods for detecting, updating, and visualizing model beliefs. arXiv preprint arXiv:2111.13654 , 2021. - · arXiv
  11. [13][13] Peter Hase, Mohit Bansal, Been Kim, and Asma Ghandeharioun. Does localization inform editing? surprising differences in causality-based localization vs. knowledge editing in language models. arXiv preprint arXiv:2301.04213 , 2023. - · arXiv
  12. [14][14] Thomas Henn, Yasukazu Sakamoto, Clément Jacquet, Shunsuke Yoshizawa, Masamichi Andou, Stephen Tchen, Ryosuke Saga, Hiroyuki Ishihara, Katsuhiko Shimizu, Yingzhen Li, et al. A principled approach to failure analysis and model repairment: Demonstration in medical imaging. In Medical Image Computing and Computer Assisted Intervention-MICCAI 2021: 24th International Conference, Strasbourg, France, September 27-October 1, 2021, Proceedings, Part III 24 , pages 509-518. Springer, 2021. -
  13. [15][15] Edward J Hu, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low-rank adaptation of large language models. In International Conference on Learning Representations , 2022. -
  14. [16][16] Zeyu Huang, Yikang Shen, Xiaofeng Zhang, Jie Zhou, Wenge Rong, and Zhang Xiong. Transformer-patcher: One mistake worth one neuron. In International Conference on Learning Representations , 2023. -
  15. [17][17] Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. Survey of hallucination in natural language generation. ACM Computing Surveys , 55(12):1-38, 2023. -
  16. [18][18] Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 , 2014. - · arXiv
  17. [19][19] James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, et al. Overcoming catastrophic forgetting in neural networks. Proceedings of the national academy of sciences , 114(13):3521-3526, 2017. -
  18. [20][20] Tom Kwiatkowski, Jennimaria Palomaki, Olivia Redfield, Michael Collins, Ankur Parikh, Chris Alberti, Danielle Epstein, Illia Polosukhin, Matthew Kelcey, Jacob Devlin, Kenton Lee, Kristina N. Toutanova, Llion Jones, Ming-Wei Chang, Andrew Dai, Jakob Uszkoreit, Quoc Le, and Slav Petrov. Natural questions: a benchmark for question answering research. Transactions of the Association of Computational Linguistics , 2019. -
  19. [21][21] Angeliki Lazaridou, Adhi Kuncoro, Elena Gribovskaya, Devang Agrawal, Adam Liska, Tayfun Terzi, Mai Gimenez, Cyprien de Masson d'Autume, Tomas Kocisky, Sebastian Ruder, et al. Mind the gap: Assessing temporal generalization in neural language models. Advances in Neural Information Processing Systems , 34:29348-29363, 2021. -
  20. [22][22] Cheolhyoung Lee, Kyunghyun Cho, and Wanmo Kang. Mixout: Effective regularization to finetune large-scale pretrained language models. In International Conference on Learning Representations , 2020. -
  21. [23][23] Yoonho Lee, Annie S Chen, Fahim Tajwar, Ananya Kumar, Huaxiu Yao, Percy Liang, and Chelsea Finn. Surgical fine-tuning improves adaptation to distribution shifts. arXiv preprint arXiv:2210.11466 , 2022. - · arXiv
  22. [25][25] Bill Yuchen Lin, Sida I Wang, Xi Lin, Robin Jia, Lin Xiao, Xiang Ren, and Scott Yih. On continual model refinement in out-of-distribution data streams. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages 3128-3139, 2022. -
  23. [26][26] Zhenhua Liu, Yunhe Wang, Kai Han, Wei Zhang, Siwei Ma, and Wen Gao. Post-training quantization for vision transformer. Advances in Neural Information Processing Systems , 34: 28092-28103, 2021. -
  24. [28][28] Kevin Meng, David Bau, Alex Andonian, and Yonatan Belinkov. Locating and editing factual associations in gpt. In Advances in Neural Information Processing Systems , 2022. -
  25. [29][29] Kevin Meng, Arnab Sen Sharma, Alex Andonian, Yonatan Belinkov, and David Bau. Massediting memory in a transformer. arXiv preprint arXiv:2210.07229 , 2023. - · arXiv
  26. [30][30] Eric Mitchell, Charles Lin, Antoine Bosselut, Chelsea Finn, and Christopher D Manning. Fast model editing at scale. In International Conference on Learning Representations , 2022. -
  27. [31][31] Eric Mitchell, Charles Lin, Antoine Bosselut, Christopher D Manning, and Chelsea Finn. Memory-based model editing at scale. In International Conference on Machine Learning , pages 15817-15831. PMLR, 2022. -
  28. [32][32] Jesse Mu, Xiang Lisa Li, and Noah Goodman. Learning to compress prompts with gist tokens. arXiv preprint arXiv:2304.08467 , 2023. - · arXiv
  29. [33][33] Jack W Rae, Sebastian Borgeaud, Trevor Cai, Katie Millican, Jordan Hoffmann, Francis Song, John Aslanides, Sarah Henderson, Roman Ring, Susannah Young, et al. Scaling language models: Methods, analysis & insights from training gopher. arXiv preprint arXiv:2112.11446 , 2021. - · arXiv
  30. [34][34] Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125 , 2022. - · arXiv
  31. [35][35] Adam Roberts, Colin Raffel, and Noam Shazeer. How much knowledge can you pack into the parameters of a language model? In Empirical Methods in Natural Language Processing (EMNLP) , 2020. -
  32. [36][36] David Rolnick, Arun Ahuja, Jonathan Schwarz, Timothy Lillicrap, and Gregory Wayne. Experience replay for continual learning. Advances in Neural Information Processing Systems , 32, 2019. -
  33. [37][37] Anton Sinitsin, Vsevolod Plokhotnyuk, Dmitry Pyrkin, Sergei Popov, and Artem Babenko. Editable neural networks. In International Conference on Learning Representations , 2019. -
  34. [39][39] Yusheng Su, Xiaozhi Wang, Yujia Qin, Chi-Min Chan, Yankai Lin, Huadong Wang, Kaiyue Wen, Zhiyuan Liu, Peng Li, Juanzi Li, et al. On transferability of prompt tuning for natural language processing. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies , pages 3949-3969, 2022. -
  35. [41][41] Frederik Träuble, Anirudh Goyal, Nasim Rahaman, Michael Mozer, Kenji Kawaguchi, Yoshua Bengio, and Bernhard Schölkopf. Discrete key-value bottleneck. arXiv preprint arXiv:2207.11240 , 2022. - · arXiv
  36. [42][42] Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in neural information processing systems , 30, 2017. -
  37. [43][43] Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Olshausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization. In International Conference on Learning Representations , 2021. -
  38. [44][44] Zifeng Wang, Zizhao Zhang, Sayna Ebrahimi, Ruoxi Sun, Han Zhang, Chen-Yu Lee, Xiaoqi Ren, Guolong Su, Vincent Perot, Jennifer Dy, et al. Dualprompt: Complementary prompting for rehearsal-free continual learning. arXiv preprint arXiv:2204.04799 , 2022. - · arXiv
  39. [45][45] Zifeng Wang, Zizhao Zhang, Chen-Yu Lee, Han Zhang, Ruoxi Sun, Xiaoqi Ren, Guolong Su, Vincent Perot, Jennifer Dy, and Tomas Pfister. Learning to prompt for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 139-149, 2022. -
  40. [46][46] Jason Weston, Sumit Chopra, and Antoine Bordes. Memory networks. arXiv preprint arXiv:1410.3916 , 2014. - · arXiv
  41. [47][47] Lee Xiong, Chenyan Xiong, Ye Li, Kwok-Fung Tang, Jialin Liu, Paul N Bennett, Junaid Ahmed, and Arnold Overwijk. Approximate nearest neighbor negative contrastive learning for dense text retrieval. In International Conference on Learning Representations , 2021. -
  42. [48][48] Yang Xu, Yutai Hou, and Wanxiang Che. Language anisotropic cross-lingual model editing. arXiv preprint arXiv:2205.12677 , 2022. - · arXiv
  43. [49][49] Chiyuan Zhang, Samy Bengio, Moritz Hardt, Benjamin Recht, and Oriol Vinyals. Understanding deep learning (still) requires rethinking generalization. Communications of the ACM , 64(3): 107-115, 2021. -
  44. [50][50] Qihuang Zhong, Liang Ding, Juhua Liu, Bo Du, and Dacheng Tao. Panda: Prompt transfer meets knowledge distillation for efficient model adaptation. arXiv preprint arXiv:2208.10160 , 2022. - · arXiv
  45. [51][51] Chen Zhu, Ankit Singh Rawat, Manzil Zaheer, Srinadh Bhojanapalli, Daliang Li, Felix Yu, and Sanjiv Kumar. Modifying memories in transformer models. arXiv preprint arXiv:2012.00363 , 2020. · arXiv

Cited by

Sorted by citations

Related

Metadata

Authors
Hartvigsen, T., Sankaranarayanan, S., Palangi, H., Kim, Y., & Ghassemi, M.
ORCID
Venue
Advances in Neural Information Processing Systems 36
Publisher
Neural Information Processing Systems Foundation, Inc. (NeurIPS)
Published
2023
Pages
47934-47959
ISSN
1049-5258 (ISSN-L)
DOI
10.52202/075280-2079
OpenAlex
W7133234277
Cited by
35
References
0
Citations by year
2026: 15 · 2025: 9 · 2024: 8 · 2023: 3
FWCI
6.21
Type
conference paper
Language
en
Open access
Locations
  • neural information processing systems (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University) (cc-by)
Funders
Canadian Institute for Advanced Research
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
Keywords
lifelong learning, model editing, continual learning, parameter-efficient fine-tuning, large language models

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