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

By Thomas Hartvigsen, S. Sankaranarayanan, Hamid Palangi, Yoon Kim, Marzyeh Ghassemi. Neural Information Processing Systems, 2022

315 citations recorded by the source corpus.

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}.

Subfields: Artificial Intelligence, Software

View Paper: https://www.semanticscholar.org/paper/560b1bc012588731b26748e33236570df777baa0

Citations in the corpus

• A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions (2023, unknown): “In comparison to the whole model, there are various techniques that involve incorporating additional parameter layers ( e.g. adapter layers (Hartvigsen et al., 2022)) as plug-ins into the original model.”
• Editing Large Language Models: Problems, Methods, and Opportunities (2023, unknown): “Differently, GRACE (Hartvigsen et al., 2022) maintains a discrete code-book as an Adapter, adding and updating elements over time to edit a model’s predictions.”