PairUni: Pairwise Training for Unified Multimodal Language Models

• URL: http://arxiv.org/abs/2510.25682v2
• Date: Thu, 30 Oct 2025 14:28:46 GMT
• Title: PairUni: Pairwise Training for Unified Multimodal Language Models
• Abstract summary: Unified vision-language models (UVLMs) must perform both understanding and generation within a single architecture.<n>We propose PairUni, a unified framework that reorganizes data into understanding-generation (UG) pairs.<n>Our approach achieves balanced improvements on various UVLMs, outperforming strong UVLM RL baselines.
• Score: 50.01064884121653
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Unified vision-language models (UVLMs) must perform both understanding and generation within a single architecture, but these tasks rely on heterogeneous data and supervision, making it difficult to balance them during reinforcement learning (RL). We propose PairUni, a unified framework that reorganizes data into understanding-generation (UG) pairs and aligns optimization accordingly. We first use GPT-o3 to augment single-task data, generating captions for understanding samples and question-answer (QA) pairs for generation samples, forming aligned pairs from the same instance. Additionally, for each generation sample, we retrieve a semantically related understanding example to form a retrieved pair, linking different but related data points. These paired structures expose cross-task semantic correspondences and support consistent policy learning. To leverage this structure, we present Pair-GPRO, a pair-aware variant based on Group Relative Policy Optimization. It assigns a similarity score to each pair to modulate the advantage, strengthening learning from well-aligned examples and reducing task interference. We curate a high-quality dataset of 16K UG pairs named PairUG for RL fine-tuning and evaluate PairUni on the powerful Janus-Pro UVLMs. Our approach achieves balanced improvements on various UVLMs, outperforming strong UVLM RL baselines. Codes are available at https://github.com/Haochen-Wang409/PairUni.

Related papers

• UniFed-VLM: Federated Instruction Tuning for Vision-Language Models with Multiple Heterogeneity [82.84279796262818]
  We propose UniFed-VLM, a unified instruction tuning framework for Vision-Language Models (VLMs)<n>It consists of two key components: 1) Federated Compensated Subspace Aggregation (FedCSA), which performs subspace-aligned aggregation of parameter-efficient adapters.<n>We conduct experiments on multiple benchmark datasets, and the results show that UniFed-VLM achieves stronger average performance.
  arXiv  Detail & Related papers  (2026-08-16T04:01:04Z)
• Improve Contrastive Clustering Performance by Multiple Fusing-Augmenting ViT Blocks [9.66175383697435]
  We design a novel fusing-augmenting ViT blocks (MFAVBs) based on the excellent feature learning ability of Vision Transformers (ViT)<n>MFAVBs serving as the backbone for contrastive clustering outperforms the state-of-the-art techniques in terms of clustering performance.
  arXiv  Detail & Related papers  (2025-11-12T01:37:48Z)
• Better Together: Leveraging Unpaired Multimodal Data for Stronger Unimodal Models [63.032359320629105]
  We introduce: Unpaired Multimodal, a modality-agnostic training paradigm in which a single model alternately processes inputs from different modalities while sharing parameters across them.<n>We show that using unpaired data from auxiliary modalities consistently improves downstream performance across diverse unimodal targets such as image and audio.
  arXiv  Detail & Related papers  (2025-10-09T17:32:23Z)
• SUDER: Self-Improving Unified Large Multimodal Models for Understanding and Generation with Dual Self-Rewards [55.99492656542475]
  We propose textbfSUDER (textbfSelf-improving textbfUnified LMMs with textbfDual stextbfElf-textbfRewards), a framework reinforcing the understanding and generation capabilities of LMMs.
  arXiv  Detail & Related papers  (2025-06-09T17:38:45Z)
• Semantic-Aligned Learning with Collaborative Refinement for Unsupervised VI-ReID [82.12123628480371]
  Unsupervised person re-identification (USL-VI-ReID) seeks to match pedestrian images of the same individual across different modalities without human annotations for model learning.<n>Previous methods unify pseudo-labels of cross-modality images through label association algorithms and then design contrastive learning framework for global feature learning.<n>We propose a Semantic-Aligned Learning with Collaborative Refinement (SALCR) framework, which builds up objective for specific fine-grained patterns emphasized by each modality.
  arXiv  Detail & Related papers  (2025-04-27T13:58:12Z)
• A Framework for Fine-Tuning LLMs using Heterogeneous Feedback [69.51729152929413]
  We present a framework for fine-tuning large language models (LLMs) using heterogeneous feedback. First, we combine the heterogeneous feedback data into a single supervision format, compatible with methods like SFT and RLHF. Next, given this unified feedback dataset, we extract a high-quality and diverse subset to obtain performance increases.
  arXiv  Detail & Related papers  (2024-08-05T23:20:32Z)
• One for all: A novel Dual-space Co-training baseline for Large-scale Multi-View Clustering [42.92751228313385]
  We propose a novel multi-view clustering model, named Dual-space Co-training Large-scale Multi-view Clustering (DSCMC) The main objective of our approach is to enhance the clustering performance by leveraging co-training in two distinct spaces. Our algorithm has an approximate linear computational complexity, which guarantees its successful application on large-scale datasets.
  arXiv  Detail & Related papers  (2024-01-28T16:30:13Z)
• Contrastive Transformer Learning with Proximity Data Generation for Text-Based Person Search [60.626459715780605]
  Given a descriptive text query, text-based person search aims to retrieve the best-matched target person from an image gallery. Such a cross-modal retrieval task is quite challenging due to significant modality gap, fine-grained differences and insufficiency of annotated data. In this paper, we propose a simple yet effective dual Transformer model for text-based person search.
  arXiv  Detail & Related papers  (2023-11-15T16:26:49Z)
• Con$^{2}$DA: Simplifying Semi-supervised Domain Adaptation by Learning Consistent and Contrastive Feature Representations [1.2891210250935146]
  Con$2$DA is a framework that extends recent advances in semi-supervised learning to the semi-supervised domain adaptation problem. Our framework generates pairs of associated samples by performing data transformations to a given input. We use different loss functions to enforce consistency between the feature representations of associated data pairs of samples.
  arXiv  Detail & Related papers  (2022-04-04T15:05:45Z)
• Learning to Rank Question Answer Pairs with Bilateral Contrastive Data Augmentation [39.22166065525888]
  We propose a novel and easy-to-apply data augmentation strategy, namely Bilateral Generation (BiG) With the augmented dataset, we design a contrastive training objective for learning to rank question answer pairs. Experimental results on three benchmark datasets, namely TREC-QA, WikiQA, and ANTIQUE, show that our method significantly improves the performance of ranking models.
  arXiv  Detail & Related papers  (2021-06-21T13:29:43Z)
• Generating Diverse and Consistent QA pairs from Contexts with Information-Maximizing Hierarchical Conditional VAEs [62.71505254770827]
  We propose a conditional variational autoencoder (HCVAE) for generating QA pairs given unstructured texts as contexts. Our model obtains impressive performance gains over all baselines on both tasks, using only a fraction of data for training.
  arXiv  Detail & Related papers  (2020-05-28T08:26:06Z)

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.