Prescribing the Right Remedy: Mitigating Hallucinations in Large Vision-Language Models via Targeted Instruction Tuning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909635353509888 |
|---|---|
| author | Hu, Rui Tu, Yahan Wei, Shuyu Lu, Dongyuan Sang, Jitao |
| author_facet | Hu, Rui Tu, Yahan Wei, Shuyu Lu, Dongyuan Sang, Jitao |
| contents | Despite achieving outstanding performance on various cross-modal tasks, current large vision-language models (LVLMs) still suffer from hallucination issues, manifesting as inconsistencies between their generated responses and the corresponding images. Prior research has implicated that the low quality of instruction data, particularly the skewed balance between positive and negative samples, is a significant contributor to model hallucinations. Recently, researchers have proposed high-quality instruction datasets, such as LRV-Instruction, to mitigate model hallucination. Nonetheless, our investigation reveals that hallucinatory concepts from different LVLMs exhibit specificity, i.e. the distribution of hallucinatory concepts varies significantly across models. Existing datasets did not consider the hallucination specificity of different models in the design processes, thereby diminishing their efficacy in mitigating model hallucination. In this paper, we propose a targeted instruction data generation framework named DFTG that tailored to the hallucination specificity of different models. Concretely, DFTG consists of two stages: hallucination diagnosis, which extracts the necessary information from the model's responses and images for hallucination diagnosis; and targeted data generation, which generates targeted instruction data based on diagnostic results. The experimental results on hallucination benchmarks demonstrate that the targeted instruction data generated by our method are more effective in mitigating hallucinations compared to previous datasets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_10332 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Prescribing the Right Remedy: Mitigating Hallucinations in Large Vision-Language Models via Targeted Instruction Tuning Hu, Rui Tu, Yahan Wei, Shuyu Lu, Dongyuan Sang, Jitao Computer Vision and Pattern Recognition Artificial Intelligence Despite achieving outstanding performance on various cross-modal tasks, current large vision-language models (LVLMs) still suffer from hallucination issues, manifesting as inconsistencies between their generated responses and the corresponding images. Prior research has implicated that the low quality of instruction data, particularly the skewed balance between positive and negative samples, is a significant contributor to model hallucinations. Recently, researchers have proposed high-quality instruction datasets, such as LRV-Instruction, to mitigate model hallucination. Nonetheless, our investigation reveals that hallucinatory concepts from different LVLMs exhibit specificity, i.e. the distribution of hallucinatory concepts varies significantly across models. Existing datasets did not consider the hallucination specificity of different models in the design processes, thereby diminishing their efficacy in mitigating model hallucination. In this paper, we propose a targeted instruction data generation framework named DFTG that tailored to the hallucination specificity of different models. Concretely, DFTG consists of two stages: hallucination diagnosis, which extracts the necessary information from the model's responses and images for hallucination diagnosis; and targeted data generation, which generates targeted instruction data based on diagnostic results. The experimental results on hallucination benchmarks demonstrate that the targeted instruction data generated by our method are more effective in mitigating hallucinations compared to previous datasets. |
| title | Prescribing the Right Remedy: Mitigating Hallucinations in Large Vision-Language Models via Targeted Instruction Tuning |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2404.10332 |