Piculet: Specialized Models-Guided Hallucination Decrease for MultiModal Large Language Models

TL;DR

Piculet is a training-free approach that reduces hallucinations in multimodal large language models by using specialized models to extract and incorporate detailed visual descriptions into the input, improving alignment without retraining.

Contribution

Introducing Piculet, a universal, training-free method that enhances MLLMs' input with visual descriptions from specialized models to decrease hallucinations.

Findings

Significantly reduces hallucinations in MLLMs.

Effective across different MLLMs without retraining.

Improves alignment between image content and generated text.

Abstract

Multimodal Large Language Models (MLLMs) have made significant progress in bridging the gap between visual and language modalities. However, hallucinations in MLLMs, where the generated text does not align with image content, continue to be a major challenge. Existing methods for addressing hallucinations often rely on instruction-tuning, which requires retraining the model with specific data, which increases the cost of utilizing MLLMs further. In this paper, we introduce a novel training-free method, named Piculet, for enhancing the input representation of MLLMs. Piculet leverages multiple specialized models to extract descriptions of visual information from the input image and combine these descriptions with the original image and query as input to the MLLM. We evaluate our method both quantitively and qualitatively, and the results demonstrate that Piculet greatly decreases hallucinations of MLLMs. Our method can be easily extended to different MLLMs while being universal.