Towards Unified Multimodal Editing with Enhanced Knowledge Collaboration
This work addresses the challenge of balancing reliability, generality, and locality in knowledge editing for Multimodal LLMs, which is a significant problem for researchers and developers working with MLLMs.
The paper proposes UniKE, a unified multimodal editing method for MLLMs that conceptualizes intrinsic and external knowledge as vectorized key-value memories. This approach ensures the edited MLLM maintains excellent reliability, generality, and locality simultaneously.
The swift advancement in Multimodal LLMs (MLLMs) also presents significant challenges for effective knowledge editing. Current methods, including intrinsic knowledge editing and external knowledge resorting, each possess strengths and weaknesses, struggling to balance the desired properties of reliability, generality, and locality when applied to MLLMs. In this paper, we propose UniKE, a novel multimodal editing method that establishes a unified perspective and paradigm for intrinsic knowledge editing and external knowledge resorting. Both types of knowledge are conceptualized as vectorized key-value memories, with the corresponding editing processes resembling the assimilation and accommodation phases of human cognition, conducted at the same semantic levels. Within such a unified framework, we further promote knowledge collaboration by disentangling the knowledge representations into the semantic and truthfulness spaces. Extensive experiments validate the effectiveness of our method, which ensures that the post-edit MLLM simultaneously maintains excellent reliability, generality, and locality. The code for UniKE is available at \url{https://github.com/beepkh/UniKE}.