DIEKAE: Difference Injection for Efficient Knowledge Augmentation and Editing of Large Language Models
This work addresses the problem of updating and augmenting knowledge in large language models for users needing current or specific information, representing an incremental improvement in efficiency.
The paper tackles the limitations of parametric knowledge in pretrained language models, such as outdated information, by introducing DIEKAE, a method that decouples knowledge processing using encoders to inject external knowledge, resulting in reduced computational costs and improved performance, with demonstrated efficiency gains over baselines in training and inference.
Pretrained Language Models (PLMs) store extensive knowledge within their weights, enabling them to recall vast amount of information. However, relying on this parametric knowledge brings some limitations such as outdated information or gaps in the training data. This work addresses these problems by distinguish between two separate solutions: knowledge editing and knowledge augmentation. We introduce Difference Injection for Efficient Knowledge Augmentation and Editing (DIEKÆ), a new method that decouples knowledge processing from the PLM (LLaMA2-7B, in particular) by adopting a series of encoders. These encoders handle external knowledge and inject it into the PLM layers, significantly reducing computational costs and improving performance of the PLM. We propose a novel training technique for these encoders that does not require back-propagation through the PLM, thus greatly reducing the memory and time required to train them. Our findings demonstrate how our method is faster and more efficient compared to multiple baselines in knowledge augmentation and editing during both training and inference. We have released our code and data at https://github.com/alessioGalatolo/DIEKAE.