F2RVLM: Boosting Fine-grained Fragment Retrieval for Multi-Modal Long-form Dialogue with Vision Language Model

• URL: http://arxiv.org/abs/2508.17714v2
• Date: Mon, 10 Nov 2025 03:33:57 GMT
• Title: F2RVLM: Boosting Fine-grained Fragment Retrieval for Multi-Modal Long-form Dialogue with Vision Language Model
• Authors: Hanbo Bi, Zhiqiang Yuan, Zexi Jia, Jiapei Zhang, Chongyang Li, Peixiang Luo, Ying Deng, Xiaoyue Duan, Jinchao Zhang,
• Abstract summary: Fine-grained Fragment Retrieval (FFR) is a task requiring models to locate query-relevant fragments.<n>F2RVLM is a generative retrieval model trained in a two-stage paradigm.<n>It outperforms popular Vision-Language Models in both in-domain and real-domain settings.
• Score: 18.21432204057241
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Traditional dialogue retrieval aims to select the most appropriate utterance or image from recent dialogue history. However, they often fail to meet users' actual needs for revisiting semantically coherent content scattered across long-form conversations. To fill this gap, we define the Fine-grained Fragment Retrieval (FFR) task, requiring models to locate query-relevant fragments, comprising both utterances and images, from multimodal long-form dialogues. As a foundation for FFR, we construct MLDR, the longest-turn multimodal dialogue retrieval dataset to date, averaging 25.45 turns per dialogue, with each naturally spanning three distinct topics. To evaluate generalization in real-world scenarios, we curate and annotate a WeChat-based test set comprising real-world multimodal dialogues with an average of 75.38 turns. Building on these resources, we explore existing generation-based Vision-Language Models (VLMs) on FFR and observe that they often retrieve incoherent utterance-image fragments. While optimized for generating responses from visual-textual inputs, these models lack explicit supervision to ensure semantic coherence within retrieved fragments. To this end, we propose F2RVLM, a generative retrieval model trained in a two-stage paradigm: (1) supervised fine-tuning to inject fragment-level retrieval knowledge, and (2) GRPO-based reinforcement learning with multi-objective rewards promoting semantic precision, relevance, and contextual coherence. To handle varying intra-fragment complexity, from locally dense to sparsely distributed, we introduce difficulty-aware curriculum sampling that ranks training instances by model-predicted difficulty and gradually exposes the model to harder samples. This boosts reasoning ability in long, multi-turn contexts. F2RVLM outperforms popular VLMs in both in-domain and real-domain settings, demonstrating superior retrieval performance.

Related papers

• ChatUMM: Robust Context Tracking for Conversational Interleaved Generation [44.19929499646892]
  Unified multimodal models (UMMs) have achieved remarkable progress yet remain constrained by a single-turn interaction paradigm.<n>We present ChatUMM, a conversational unified model that excels at robust context tracking to sustain interleaved multimodal generation.<n>ChatUMM derives its capabilities from an interleaved multi-turn training strategy that models serialized text-image streams as a continuous conversational flow.
  arXiv  Detail & Related papers  (2026-02-06T07:11:50Z)
• Exploring a Unified Vision-Centric Contrastive Alternatives on Multi-Modal Web Documents [99.62178668680578]
  We propose Vision-Centric Contrastive Learning (VC2L), a unified framework that models text, images, and their combinations using a single vision transformer.<n> VC2L operates entirely in pixel space by rendering all inputs, whether textual, visual, or combined, as images.<n>To capture complex cross-modal relationships in web documents, VC2L employs a snippet-level contrastive learning objective that aligns consecutive multimodal segments.
  arXiv  Detail & Related papers  (2025-10-21T14:59:29Z)
• Test-Time Scaling Strategies for Generative Retrieval in Multimodal Conversational Recommendations [70.94563079082751]
  E-commerce has exposed the limitations of traditional product retrieval systems in managing complex, multi-turn user interactions.<n>We propose a novel framework that introduces test-time scaling into conversational multimodal product retrieval.<n>Our approach builds on a generative retriever, further augmented with a test-time reranking mechanism that improves retrieval accuracy and better aligns results with evolving user intent throughout the dialogue.
  arXiv  Detail & Related papers  (2025-08-25T15:38:56Z)
• Unified Generative and Discriminative Training for Multi-modal Large Language Models [88.84491005030316]
  Generative training has enabled Vision-Language Models (VLMs) to tackle various complex tasks. Discriminative training, exemplified by models like CLIP, excels in zero-shot image-text classification and retrieval. This paper proposes a unified approach that integrates the strengths of both paradigms.
  arXiv  Detail & Related papers  (2024-11-01T01:51:31Z)
• Interactive Text-to-Image Retrieval with Large Language Models: A Plug-and-Play Approach [33.231639257323536]
  In this paper, we address the issue of dialogue-form context query within the interactive text-to-image retrieval task. By reformulating the dialogue-form context, we eliminate the necessity of fine-tuning a retrieval model on existing visual dialogue data. We construct the LLM questioner to generate non-redundant questions about the attributes of the target image.
  arXiv  Detail & Related papers  (2024-06-05T16:09:01Z)
• DialCLIP: Empowering CLIP as Multi-Modal Dialog Retriever [83.33209603041013]
  We propose a parameter-efficient prompt-tuning method named DialCLIP for multi-modal dialog retrieval. Our approach introduces a multi-modal context generator to learn context features which are distilled into prompts within the pre-trained vision-language model CLIP. To facilitate various types of retrieval, we also design multiple experts to learn mappings from CLIP outputs to multi-modal representation space.
  arXiv  Detail & Related papers  (2024-01-02T07:40:12Z)
• Collaborative Reasoning on Multi-Modal Semantic Graphs for Video-Grounded Dialogue Generation [53.87485260058957]
  We study video-grounded dialogue generation, where a response is generated based on the dialogue context and the associated video. The primary challenges of this task lie in (1) the difficulty of integrating video data into pre-trained language models (PLMs) We propose a multi-agent reinforcement learning method to collaboratively perform reasoning on different modalities.
  arXiv  Detail & Related papers  (2022-10-22T14:45:29Z)
• An Exploratory Study on Long Dialogue Summarization: What Works and What's Next [33.1899354772074]
  We study long dialogue summarization by investigating three strategies to deal with the lengthy input problem and locate relevant information. Our experimental results on three long dialogue datasets (QMSum, MediaSum, SummScreen) show that the retrieve-then-summarize pipeline models yield the best performance.
  arXiv  Detail & Related papers  (2021-09-10T01:38:26Z)
• Video-Grounded Dialogues with Pretrained Generation Language Models [88.15419265622748]
  We leverage the power of pre-trained language models for improving video-grounded dialogue. We propose a framework by formulating sequence-to-grounded dialogue tasks as a sequence-to-grounded task. Our framework allows fine-tuning language models to capture dependencies across multiple modalities.
  arXiv  Detail & Related papers  (2020-06-27T08:24:26Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.