Salvato in:
Dettagli Bibliografici
Autori principali: Yi-Ge, Ellen, Gao, Jiechao, Han, Wei, Zhu, Wei
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2412.07619
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913606029803520
author Yi-Ge, Ellen
Gao, Jiechao
Han, Wei
Zhu, Wei
author_facet Yi-Ge, Ellen
Gao, Jiechao
Han, Wei
Zhu, Wei
contents Recently, large language models (LLMs) have demonstrated impressive capabilities in dealing with new tasks with the help of in-context learning (ICL). In the study of Large Vision-Language Models (LVLMs), when implementing ICL, researchers usually adopts the naive strategies like fixed demonstrations across different samples, or selecting demonstrations directly via a visual-language embedding model. These methods does not guarantee the configured demonstrations fit the need of the LVLMs. To address this issue, we now propose a novel framework, \underline{d}emonstration \underline{r}etriever for large m\underline{u}lti-modal \underline{m}odel (DRUM), which fine-tunes the visual-language embedding model to better meet the LVLM's needs. First, we discuss the retrieval strategies for a visual-language task, assuming an embedding model is given. And we propose to concate the image and text embeddings to enhance the retrieval performance. Second, we propose to re-rank the demonstrations retrieved by the embedding model via the LVLM's feedbacks, and calculate a list-wise ranking loss for training the embedding model. Third, we propose an iterative demonstration mining strategy to improve the training of the embedding model. Through extensive experiments on 3 types of visual-language tasks, 7 benchmark datasets, our DRUM framework is proven to be effective in boosting the LVLM's in-context learning performance via retrieving more proper demonstrations.
format Preprint
id arxiv_https___arxiv_org_abs_2412_07619
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DRUM: Learning Demonstration Retriever for Large MUlti-modal Models
Yi-Ge, Ellen
Gao, Jiechao
Han, Wei
Zhu, Wei
Computation and Language
Recently, large language models (LLMs) have demonstrated impressive capabilities in dealing with new tasks with the help of in-context learning (ICL). In the study of Large Vision-Language Models (LVLMs), when implementing ICL, researchers usually adopts the naive strategies like fixed demonstrations across different samples, or selecting demonstrations directly via a visual-language embedding model. These methods does not guarantee the configured demonstrations fit the need of the LVLMs. To address this issue, we now propose a novel framework, \underline{d}emonstration \underline{r}etriever for large m\underline{u}lti-modal \underline{m}odel (DRUM), which fine-tunes the visual-language embedding model to better meet the LVLM's needs. First, we discuss the retrieval strategies for a visual-language task, assuming an embedding model is given. And we propose to concate the image and text embeddings to enhance the retrieval performance. Second, we propose to re-rank the demonstrations retrieved by the embedding model via the LVLM's feedbacks, and calculate a list-wise ranking loss for training the embedding model. Third, we propose an iterative demonstration mining strategy to improve the training of the embedding model. Through extensive experiments on 3 types of visual-language tasks, 7 benchmark datasets, our DRUM framework is proven to be effective in boosting the LVLM's in-context learning performance via retrieving more proper demonstrations.
title DRUM: Learning Demonstration Retriever for Large MUlti-modal Models
topic Computation and Language
url https://arxiv.org/abs/2412.07619