Augmenting Multimodal LLMs with Self-Reflective Tokens for Knowledge-based Visual Question Answering
This work addresses the problem of limited adaptability in MLLMs for practical applications like visual question answering, representing an incremental improvement by enhancing existing models with external knowledge integration.
The paper tackles the limitation of multimodal LLMs (MLLMs) in handling tasks requiring external knowledge by introducing Reflective LLaVA, which uses reflective tokens to dynamically integrate external knowledge sources, resulting in superior performance for knowledge-based visual question answering compared to existing methods.
Multimodal LLMs (MLLMs) are the natural extension of large language models to handle multimodal inputs, combining text and image data. They have recently garnered attention due to their capability to address complex tasks involving both modalities. However, their effectiveness is limited to the knowledge acquired during training, which restricts their practical utility. In this work, we introduce a novel method to enhance the adaptability of MLLMs by integrating external knowledge sources. Our proposed model, Reflective LLaVA (ReflectiVA), utilizes reflective tokens to dynamically determine the need for external knowledge and predict the relevance of information retrieved from an external database. Tokens are trained following a two-stage two-model training recipe. This ultimately enables the MLLM to manage external knowledge while preserving fluency and performance on tasks where external knowledge is not needed. Through our experiments, we demonstrate the efficacy of ReflectiVA for knowledge-based visual question answering, highlighting its superior performance compared to existing methods. Source code and trained models are publicly available at https://aimagelab.github.io/ReflectiVA.