English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Large Transformer-based Pretrained Language Models (PLMs) dominate almost allNatural Language Processing (NLP) tasks. Nevertheless, they still make mistakesfrom time to time. For a model deployed in an industrial environment, fixingthese mistakes quickly and robustly is vital to improve user experiences.Previous works formalize such problems as Model Editing (ME) and mostly focuson fixing one mistake. However, the one-mistake-fixing scenario is not anaccurate abstraction of the real-world challenge. In the deployment of AIservices, there are ever-emerging mistakes, and the same mistake may recur ifnot corrected in time. Thus a preferable solution is to rectify the mistakes assoon as they appear nonstop. Therefore, we extend the existing ME intoSequential Model Editing (SME) to help develop more practical editing methods.Our study shows that most current ME methods could yield unsatisfying resultsin this scenario. We then introduce Transformer-Patcher, a novel model editorthat can shift the behavior of transformer-based models by simply adding andtraining a few neurons in the last Feed-Forward Network layer. Experimentalresults on both classification and generation tasks show thatTransformer-Patcher can successively correct up to thousands of errors(Reliability) and generalize to their equivalent inputs (Generality) whileretaining the model's accuracy on irrelevant inputs (Locality). Our methodoutperforms previous fine-tuning and HyperNetwork-based methods and achievesstate-of-the-art performance for Sequential Model Editing (SME). The code isavailable at https://github.com/ZeroYuHuang/Transformer-Patcher.

Transformer-Patcher: One Mistake worth One Neuron