Multimodal Large Language Models (MLLMs) are often constrained by a language-space bottleneck, forcing complex visual reasoning into discrete tokens which can lose perceptual nuance. A promising alternative is continuous latent reasoning, where the goal is to discover implicit reasoning pathways that bridge the multimodal query and the final answer. However, this introduces a severe train-inference mismatch: a training-time posterior, conditioned on the ground-truth answer, can exploit answer-dependent shortcuts. Standard variational training then forces the inference-time prior to mimic a posterior that has access to information unavailable at test time, leading to poor performance. To address this, we propose Asymmetric Mutual Variational Learning (AMVL), a framework that resolves this mismatch via a bidirectional calibration objective. A forward KL divergence trains the target-agnostic prior to match the posterior, while a novel reverse KL divergence simultaneously regularizes the posterior, preventing it from collapsing into inference-incompatible regions and mitigating this ``answer leakage''. We provide theoretical analysis formalizing this leakage as prior contamination and prove that our dual-KL objective reduces it. We instantiate AMVL in a latent-integrated MLLM and show that it consistently outperforms strong discrete and latent-reasoning baselines, improving the average score on the complex BLINK benchmark by +10.83 and achieving gains of up to +32.00 on individual reasoning tasks, with analyses confirming improved latent-space stability.
M-MRE: Extending the Mutual Reinforcement Effect to Multimodal Information Extraction
Yokohama National University(Yokohama国立大学);Monash University(墨尔本大学);Southern University of Science and Technology(南方科技大学);Hong Kong University of Science and Technology(香港科学大学);University of Chinese Academy of Sciences(中国科学院大学);Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences(中国科学院深圳先进技术研究所)
Unsupervised Domain Adaptation on Person Re-Identification via Dual-level Asymmetric Mutual Learning
Xiamen University(厦门大学);China University of Mining and Technology(中国矿业大学);Peng Cheng Laboratory(鹏城实验室);University of Chinese Academy of Sciences(中国科学院大学);Tencent(腾讯)
StructXLIP: Enhancing Vision-language Models with Multimodal Structural Cues
University of Verona(维罗纳大学);Sapienza University of Rome(罗马大学);University of Trento(特伦托大学);Fondazione Bruno Kessler(布鲁诺·凯斯勒基金会);Reykjavik University(雷克雅未克大学)
GraphMMP: A Graph Neural Network Model with Mutual Information and Global Fusion for Multimodal Medical Prognosis
Hangzhou Dianzi University(杭州电子科技大学);Shenzhen Polytechnic University(深圳职业技术大学);The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));Shenzhen Research Institute of Big Data(深圳大数据研究院);Shenzhen Institutes of Advanced Technology(深圳先进技术研究院);Yunnan University(云南大学);Shenzhen University(深圳大学)
Asymmetric Reinforcing against Multi-modal Representation Bias
College of Intelligence and Computing, Tianjin University(天津大学智能与计算学部);The State Key Laboratory of Integrated Services Networks, Xidian University(西安电子科技大学综合业务网理论及关键技术国家重点实验室)
Multimodal Representation Learning via Maximization of Local Mutual Information
CSAIL, Massachusetts Institute of Technology(麻省理工学院计算机科学与人工智能实验室);MIT Lincoln Laboratory(麻省理工学院林肯实验室);Beth Israel Deaconess Medical Center, Harvard Medical School(哈佛医学院贝斯以色列女执事医疗中心)
Robo-MUTUAL: Robotic Multimodal Task Specification via Unimodal Learning
Institute for AI Industry Research (AIR), Tsinghua University(清华大学人工智能产业研究院(AIR));College of Engineering, Peking University(北京大学工学院);Sensetime Research(商汤科技研究院);University of Toronto(多伦多大学);University College London(伦敦大学学院);Shanghai AI Lab(上海人工智能实验室)
Multimodal Fake News Detection: MFND Dataset and Shallow-Deep Multitask Learning
School of Artificial Intelligence, Hebei University of Technology(河北工业大学人工智能学院);School of Computing and Information Technology, Great Bay University(大湾大学计算与信息学院);Guangdong Provincial Key Laboratory of Intelligent Information Processing & Shenzhen Key Laboratory of Media Security, Shenzhen University(广东省智能信息处理重点实验室 & 深圳市媒体安全重点实验室);Dongguan Key Laboratory for Intelligence and Information Technology(东莞智能信息科技重点实验室)
MDD-Net: Multimodal Depression Detection through Mutual Transformer
Centre for Pattern Analysis and Machine Intelligence, Department of Electrical and Computer Engineering, University of Waterloo(模式分析与机器智能中心,电气与计算机工程系,滑铁卢大学);School of Engineering and Computing, Department of Computer Science and Engineering, American University of Ras Al Khaimah(工程与计算学院,计算机科学与工程系,阿联酋拉线哈姆市美国大学)
GeoBalance: Geometry-Aware Monitoring and Reconstruction with Asymmetric Optimization for Balanced Multimodal Learning
Beihang University(北京航空航天大学);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);Amap, Alibaba Group(阿里巴巴集团高德);Kuaishou Technology(快手科技);Beijing Jiaotong University(北京交通大学);Renmin University of China(中国人民大学)
Traits Run Deeper: Trait-Specific Asymmetric Fusion for Multimodal Personality Assessment
Hefei University of Technology(合肥工业大学);Intelligent Interconnected Systems Laboratory of Anhui Province(安徽省智能互联系统实验室);Jianghuai Advanced Technology Center(江淮前沿技术中心);Anhui Provincial Industry Innovation Center of Humanoid Robots(安徽省人形机器人产业创新中心);Anhui Provincial Key Laboratory of Humanoid Robots(安徽省人形机器人重点实验室)
提出Traits Run Deeper框架,通过多模态基础表示、特质特异性非对称融合和分布校准回归模块,解决人格评估中模态偏好差异和标签偏差问题,在AVI Challenge 2026上MSE降低约25%。