arXiv AI

TTE-Flash: Accelerating Reasoning-based Multimodal Representations via Think-Then-Embed Tokens

arXiv Computer Vision
Sep 1

Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings

arXiv:2604.22280v4 Announce Type: replace Abstract: Multimodal Large Language Models (MLLMs) have emerged as a promising foundation for universal multimodal embeddings. Recent studies have shown that...

By Peixi Wu, Ke Mei, Feipeng Ma, Bosong Chai, Zhibin Lan, Chenxi Zhao, Shannan Yan, Jie Chen, Zhangchi Hu, Yansong Peng, Bo Lin, Junjie Zhou, Dacheng Yin, Tianyi Wang, Fengyun Rao, Jing Lyu, Hebei Li, Xiaoyan Sun
arXiv Computation and Language
Aug 27

MMEmb-R1: Reasoning-Enhanced Multimodal Embedding with Pair-Aware Selection and Adaptive Control

MMEmb-R1 is a multimodal embedding framework that enhances reasoning by treating it as a latent variable and selecting beneficial reasoning paths through pair-aware selection and counterfactual intervention. It uses reinforcement learning to invoke reasoning only when necessary, reducing unnecessary computation and latency. On the MMEB-V2 benchmark, MMEmb-R1 achieves a state‑of‑the‑art score of 71.2 with just 4 B parameters.

By Yuchi Wang, Dingkang Yang, Haiyang Yu, Weikang Bian, Jiefeng Long, Xiao Liang, Chao Feng, Hongsheng Li
arXiv Computer Vision
Oct 12

VaME: Exploring Variational Latent Reasoning for Multimodal Embeddings

VaME (Variational Multimodal Embeddings) introduces a framework that treats latent reasoning as a learnable distribution over trajectories, enabling autoregressive exploration in latent space guided by answer reconstruction. By augmenting embedding-token readout with a latent-fused embedding and optimizing over stochastic variational trajectories via reinforcement learning with a Semantic Decoding Reward, VaME achieves superior performance on the MMEB-V2 benchmark and reasoning-intensive tasks like MRMR. The method also delivers at least a 4.25× inference speedup compared to deterministic latent autoregressive baselines.

By Peixi Wu, Mingzhou Jiang, Feipeng Ma, Biao Yang, Yunhao Zhou, Wei Yuan, Bosong Chai, Huizu Lin, Jie Chen, Zhangchi Hu, Fan Yang, Wenwu Ou, Hebei Li, Xiaoyan Sun
arXiv AI
Sep 28

Skip the Talk, Re-Focus on Vision: Latent Reasoning for Reasoning Segmentation in Multimodal Large Language Models

The paper introduces LIRSeg, a method that replaces explicit Chain-of-Thought reasoning in multimodal large language models with a compact set of learnable latent tokens for reasoning segmentation. LIRSeg is trained in two stages—spatial alignment and GRPO—while employing extreme-advantage sampling, decoupled exploration-stability updates, and latent diversity amplification to enhance token informativeness. Experiments show that LIRSeg improves segmentation accuracy and reasoning efficiency, achieving significant gIoU gains over the VisionReasoner baseline and reducing reasoning tokens by about 16×.

By Tianhang Guo, Yulin He, Wei Chen, Wenjuan Zhou, Yuhang Li, Xinbiao Gan
arXiv AI
Oct 26

Beyond Post-Hoc Rewards: Privileged Reasoning Exploration for Multimodal Embeddings

The paper introduces PILOT, a framework that treats reasoning as a latent variable to enhance multimodal embeddings. By modeling prior and posterior reasoning modes with a shared large language model, PILOT uses relational supervision during training to guide exploration of retrieval‑relevant reasoning traces. On the MMEB‑V2 benchmark, PILOT outperforms baselines across all task groups, notably improving visual‑document retrieval scores by over six points.

By Tianyu Yang, Yijun Tian, Lars Benedikt Kaesberg, Jan Philip Wahle, Bela Gipp, Terry Ruas, Daniel Kurzawe
arXiv Machine Learning
Sep 7

Latent-Aligned Reasoning for Multimodal Recommendation

The paper introduces LARK, a two‑stage latent reasoning framework designed to mitigate cross‑modal dilution in multimodal recommendation systems. In the first stage, learnable latent tokens are interleaved with chain‑of‑thought reasoning and aligned with a frozen vision encoder to preserve visual details. The second stage projects these latent representations through a bridge MLP, employing item‑to‑item contrastive learning and aligning intermediate features with the first‑stage hidden states to anchor final embeddings to the model’s reasoning output. Experiments on three public benchmarks and an industrial dataset demonstrate that LARK achieves state‑of‑the‑art performance across multiple recommendation architectures, with ablation studies confirming the contribution of each component.

By Jiarui Jin, Anyang Ji
arXiv AI
Sep 25

Is Reasoning Always Useful? Rethinking Reasoning Utility in Universal Multimodal Embeddings

The paper investigates whether reasoning always benefits universal multimodal embeddings (UMEs). By comparing the discriminative and reasoning-driven branches of UME-R1, the authors find that while reasoning improves positive similarity in 56.6% of cases, it also creates 15.7% false-helpful instances where hard negatives are drawn closer. Diagnostic analyses reveal that reasoning often de‑condenses retrieved neighborhoods and that chain‑of‑thought tokens encode evidence common to both positives and hard negatives. Based on these insights, the authors introduce SURE, a utility router that boosts UME-R1‑7B by 1.5 points and consistently improves other embedding models on MMEB‑V2 without retraining or extra VLM passes.

By Wenxiao Fan, Jingling Fu, Luohang Liu, Xinyuan Shan, Lichen Ma, Yu He, Junshi Huang, Yan Li, Kan Li
arXiv Computer Vision
Oct 9

Look Back, Think Ahead: Visual Memory on Demand for Efficient Multimodal Reasoning

The paper introduces ViMoD, a lightweight framework that dynamically manages visual memory for multimodal large language models. ViMoD uses Deformable Aggregation of Region-wise Tokens (DART) to create content‑adaptive coarse representations linked to recoverable fine tokens, and Temporal Routing for Adaptive Contextual Evidence (TRACE) to anticipate and adjust evidence needs during reasoning. On the Qwen3‑VL‑4B model, ViMoD achieves superior performance across eight reasoning benchmarks with only a 20% visual token budget and minimal additional parameters.

By Yicheng Xue, Han Wu, Jufeng Yang, Minjing Dong, Xinghao Chen, Hanting Chen, Jianyuan Guo