Despite recent advancements in Multi-modal Large Language Models (MLLMs) on diverse understanding tasks, these models struggle to solve problems which require extensive multi-step reasoning. This is primarily due to the progressive dilution of visual information during long-context generation, which hinders their ability to fully exploit test-time scaling. To address this issue, we introduce Vision-aligned Latent Reasoning (VaLR), a simple, yet effective reasoning framework that dynamically generates vision-aligned latent tokens before each Chain of Thought reasoning step, guiding the model to reason based on perceptual cues in the latent space. Specifically, VaLR is trained to preserve visual knowledge during reasoning by aligning intermediate embeddings of MLLM with those from vision encoders. Empirical results demonstrate that VaLR consistently outperforms existing approaches across a wide range of benchmarks requiring long-context understanding or precise visual perception, while exhibiting test-time scaling behavior not observed in prior MLLMs. In particular, VaLR improves the performance significantly from 33.0% to 52.9% on VSI-Bench, achieving a 19.9%p gain over Qwen2.5-VL.
Reasoning-Aligned Perception Decoupling for Scalable Multi-modal Reasoning
Southern University of Science and Technology（南方科技大学）；The Hong Kong University of Science and Technology（香港科技大学）；Huawei Noah’s Ark Lab（华为诺亚实验室）；Huawei Cloud Project（华为云项目）
University of California, Santa Barbara（加州大学圣巴巴拉分校）
A major challenge for video captioning is to combine audio and visual cues. Existing multi-modal fusion methods have shown encouraging results in video understanding. However, the temporal structures of multiple modalities at different granularities are rarely explored, and how to selectively fuse the multi-modal representations at different levels of details remains uncharted. In this paper, we propose a novel hierarchically aligned cross-modal attention (HACA) framework to learn and selectively fuse both global and local temporal dynamics of different modalities. Furthermore, for the first t…
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；University at Buffalo（布法罗大学）；Amazon（亚马逊）
Contrastive loss has been increasingly used in learning representations from multiple modalities. In the limit, the nature of the contrastive loss encourages modalities to exactly match each other in the latent space. Yet it remains an open question how the modality alignment affects the downstream task performance. In this paper, based on an information-theoretic argument, we first prove that exact modality alignment is sub-optimal in general for downstream prediction tasks. Hence we advocate that the key of better performance lies in meaningful latent modality structures instead of perfect m…
The Chinese University of Hong Kong（香港中文大学）；SenseTime Research（商汤科技研究院）；Tsinghua University（清华大学）
Exploiting relationships between visual regions and question words have achieved great success in learning multi-modality features for Visual Question Answering (VQA). However, we argue that existing methods mostly model relations between individual visual regions and words, which are not enough to correctly answer the question. From humans' perspective, answering a visual question requires understanding the summarizations of visual and language information. In this paper, we proposed the Multi-modality Latent Interaction module (MLI) to tackle this problem. The proposed module learns the cros…
LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning
School of AI, Shanghai Jiao Tong University（上海交通大学人工智能学院）；Kling Team, Kuaishou Technology（快手科技 Kling 团队）；Peking University（北京大学）；HKUST（香港科技大学）；CASIA（中国科学院自动化研究所）；Nanjing University（南京大学）；Renmin University of China（中国人民大学）；Tsinghua University（清华大学）
本文提出LatentOmni框架，通过统一的音频-视觉潜在空间进行多模态推理，利用特征级监督和Omni-Sync Position Embedding保持时间一致性，从而在多个音频-视觉推理基准测试中取得最佳性能。
From Consistency to Complementarity: Aligned and Disentangled Multi-modal Learning for Time Series Understanding and Reasoning
The Hong Kong University of Science（香港科学与技术大学）；Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）
University of Toronto（多伦多大学）；MIT CSAIL（麻省理工学院计算机科学与人工智能实验室）；University of Maryland - Baltimore County（马里兰大学巴尔的摩县分校）
People can recognize scenes across many different modalities beyond natural images. In this paper, we investigate how to learn cross-modal scene representations that transfer across modalities. To study this problem, we introduce a new cross-modal scene dataset. While convolutional neural networks can categorize cross-modal scenes well, they also learn an intermediate representation not aligned across modalities, which is undesirable for cross-modal transfer applications. We present methods to regularize cross-modal convolutional neural networks so that they have a shared representation that i…
School of Computer & Communication Engineering, University of Science and Technology Beijing（北京科技大学计算机与通信工程学院）
Learning a common latent embedding by aligning the latent spaces of cross-modal autoencoders is an effective strategy for Generalized Zero-Shot Classification (GZSC). However, due to the lack of fine-grained instance-wise annotations, it still easily suffer from the domain shift problem for the discrepancy between the visual representation of diversified images and the semantic representation of fixed attributes. In this paper, we propose an innovative autoencoder network by learning Aligned Cross-Modal Representations (dubbed ACMR) for GZSC. Specifically, we propose a novel Vision-Semantic Al…
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Renmin University of China（中国人民大学）；CEBSIT, Chinese Academy of Sciences（中国科学院CEBSIT）
Multispectral pedestrian detection has shown great advantages under poor illumination conditions, since the thermal modality provides complementary information for the color image. However, real multispectral data suffers from the position shift problem, i.e. the color-thermal image pairs are not strictly aligned, making one object has different positions in different modalities. In deep learning based methods, this problem makes it difficult to fuse the feature maps from both modalities and puzzles the CNN training. In this paper, we propose a novel Aligned Region CNN (AR-CNN) to handle the w…
Generative adversarial network (GAN) has achieved impressive success on cross-domain generation, but it faces difficulty in cross-modal generation due to the lack of a common distribution between heterogeneous data. Most existing methods of conditional based cross-modal GANs adopt the strategy of one-directional transfer and have achieved preliminary success on text-to-image transfer. Instead of learning the transfer between different modalities, we aim to learn a synchronous latent space representing the cross-modal common concept. A novel network component named synchronizer is proposed in t…
Michelangelo: Conditional 3D Shape Generation based on Shape-Image-Text Aligned Latent Representation
ShanghaiTech University（上海科技大学）；Tencent PCG（腾讯平台与内容事业群）；Fudan University（复旦大学）；Shanghai Engineering Research Center of Intelligent Vision and Imaging（上海智能视觉与成像工程研究中心）；Shanghai Engineering Research Center of Energy Efficient and Custom AI IC（上海高效节能定制人工智能芯片工程研究中心）
What Matters for Diffusion-Friendly Latent Manifold? Prior-Aligned Autoencoders for Latent Diffusion
Shanghai Jiao Tong University（上海交通大学）；Alibaba Group（阿里巴巴集团）；Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences（深圳先进技术研究院，中国科学院）；Beihang University（北航）；Sun Yat-sen University（中山大学）；Nankai University（南开大学）；Shanghai AI Laboratory（上海人工智能实验室）
Cross-Modal-Domain Generalization Through Semantically Aligned Discrete Representations
Peter L. Reichertz Institute for Medical Informatics（汉诺威医学院彼得·L·里赫茨医学信息学研究所）；Lower Saxony Center for AI and Causal Methods in Medicine (CAIMed)（下萨克森医学人工智能与因果方法中心（CAIMed））
Imperial College London（帝国理工学院）；Siemens Healthineers（西门子医疗）
We propose a fully unsupervised multi-modal deformable image registration method (UMDIR), which does not require any ground truth deformation fields or any aligned multi-modal image pairs during training. Multi-modal registration is a key problem in many medical image analysis applications. It is very challenging due to complicated and unknown relationships between different modalities. In this paper, we propose an unsupervised learning approach to reduce the multi-modal registration problem to a mono-modal one through image disentangling. In particular, we decompose images of both modalities …
Nanyang Technological University（南洋理工大学）；Singapore Management University（新加坡管理大学）；Salesforce Research Asia（赛富时亚洲研究院）
This paper investigates an open research problem of generating text-image pairs to improve the training of fine-grained image-to-text cross-modal retrieval task, and proposes a novel framework for paired data augmentation by uncovering the hidden semantic information of StyleGAN2 model. Specifically, we first train a StyleGAN2 model on the given dataset. We then project the real images back to the latent space of StyleGAN2 to obtain the latent codes. To make the generated images manipulatable, we further introduce a latent space alignment module to learn the alignment between StyleGAN2 latent …
Multi-view radar-camera fused 3D object detection provides a farther detection range and more helpful features for autonomous driving, especially under adverse weather. The current radar-camera fusion methods deliver kinds of designs to fuse radar information with camera data. However, these fusion approaches usually adopt the straightforward concatenation operation between multi-modal features, which ignores the semantic alignment with radar features and sufficient correlations across modals. In this paper, we present MVFusion, a novel Multi-View radar-camera Fusion method to achieve semantic…
Wangxuan Institute of Computer Technology, Peking University（北京大学王选计算机研究所）；Microsoft Research（微软研究院）
Language-guided image generation has achieved great success nowadays by using diffusion models. However, texts can be less detailed to describe highly-specific subjects such as a particular dog or a certain car, which makes pure text-to-image generation not accurate enough to satisfy user requirements. In this work, we present a novel Unified Multi-Modal Latent Diffusion (UMM-Diffusion) which takes joint texts and images containing specified subjects as input sequences and generates customized images with the subjects. To be more specific, both input texts and images are encoded into one unifi…
MM-LDM: Multi-Modal Latent Diffusion Model for Sounding Video Generation
Institute of Automation, Chinese Academy of Science（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；University of Adelaide（阿德莱德大学）
Cross-modal Offset-guided Dynamic Alignment and Fusion for Weakly Aligned UAV Object Detection
State Key Laboratory of Optical Field Manipulation Science and Technology, Chinese Academy of Sciences（光学场操控科学与技术国家重点实验室，中国科学院）；Institute of Optics and Electronics, Chinese Academy of Sciences（中国科学院光电研究所）
Self-Supervised Generative-Contrastive Learning of Multi-Modal Euclidean Input for 3D Shape Latent Representations: A Dynamic Switching Approach
Institute for Anthropomatics and Robotics（人机化与机器人研究所）；Karlsruhe Institute of Technology（卡尔斯鲁厄理工大学）；Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB（弗劳恩霍夫光学、系统技术与图像利用研究所）
Zero-Shot Video Restoration and Enhancement with Text-to-Image Latent Diffusion Models and Multi-Modal References
School of Electrical and Information Engineering, Tianjin University（天津大学电气自动化与信息工程学院）；Lappeenranta-Lahti University of Technology LUT（拉彭兰塔-拉赫蒂理工大学（LUT））
Aligned Consensus Teaching for Label-Efficient Oriented Object Detection in Weakly-Aligned Visible-Infrared Imagery
Beijing University of Technology（北京工业大学）；University of Science and Technology of China（中国科学技术大学）；Peking University（北京大学）；Institute of Software, Chinese Academy of Sciences（中国科学院软件研究所）；Beijing Institute of Technology（北京理工大学）；Fuzhou University（福州大学）；China University of Geosciences Wuhan（中国地质大学（武汉））；Ghent University（根特大学）
The human hand moves in complex and high-dimensional ways, making estimation of 3D hand pose configurations from images alone a challenging task. In this work we propose a method to learn a statistical hand model represented by a cross-modal trained latent space via a generative deep neural network. We derive an objective function from the variational lower bound of the VAE framework and jointly optimize the resulting cross-modal KL-divergence and the posterior reconstruction objective, naturally admitting a training regime that leads to a coherent latent space across multiple modalities such …
Variational autoencoder (VAE) has widely been utilized for modeling data distributions because it is theoretically elegant, easy to train, and has nice manifold representations. However, when applied to image reconstruction and synthesis tasks, VAE shows the limitation that the generated sample tends to be blurry. We observe that a similar problem, in which the generated trajectory is located between adjacent lanes, often arises in VAE-based trajectory forecasting models. To mitigate this problem, we introduce a hierarchical latent structure into the VAE-based forecasting model. Based on the a…
Multi-modal generative models represent an important family of deep models, whose goal is to facilitate representation learning on data with multiple views or modalities. However, current deep multi-modal models focus on the inference of shared representations, while neglecting the important private aspects of data within individual modalities. In this paper, we introduce a disentangled multi-modal variational autoencoder (DMVAE) that utilizes disentangled VAE strategy to separate the private and shared latent spaces of multiple modalities. We specifically consider the instance where the laten…
LEAF: Latent Diffusion with Efficient Encoder Distillation for Aligned Features in Medical Image Segmentation
School of Computer Science and Engineering, Sun Yat-sen University, China（计算机科学与工程学院，中山大学）；Department of Biomedical Engineering, Johns Hopkins University, United States（生物医学工程系，约翰霍普金斯大学）
We propose a novel approach for multi-modal Image-to-image (I2I) translation. To tackle the one-to-many relationship between input and output domains, previous works use complex training objectives to learn a latent embedding, jointly with the generator, that models the variability of the output domain. In contrast, we directly model the style variability of images, independent of the image synthesis task. Specifically, we pre-train a generic style encoder using a novel proxy task to learn an embedding of images, from arbitrary domains, into a low-dimensional style latent space. The learned la…
University of Dundee（邓迪大学）；University of Cambridge（剑桥大学）；University of Bath（巴斯大学）；East China Normal University（华东师范大学）
MRI synthesis promises to mitigate the challenge of missing MRI modality in clinical practice. Diffusion model has emerged as an effective technique for image synthesis by modelling complex and variable data distributions. However, most diffusion-based MRI synthesis models are using a single modality. As they operate in the original image domain, they are memory-intensive and less feasible for multi-modal synthesis. Moreover, they often fail to preserve the anatomical structure in MRI. Further, balancing the multiple conditions from multi-modal MRI inputs is crucial for multi-modal synthesis. …
Zhejiang University（浙江大学）；Singapore Management University（新加坡管理大学）；Alibaba-Zhejiang University Joint Institute of Frontier Technologies（阿里巴巴-浙江大学前沿技术联合研究院）
Nowadays, driven by the increasing concern on diet and health, food computing has attracted enormous attention from both industry and research community. One of the most popular research topics in this domain is Food Retrieval, due to its profound influence on health-oriented applications. In this paper, we focus on the task of cross-modal retrieval between food images and cooking recipes. We present Modality-Consistent Embedding Network (MCEN) that learns modality-invariant representations by projecting images and texts to the same embedding space. To capture the latent alignments between mod…
University of Insubria（因苏布里亚大学）；National University of Sciences and Technology（国立科技大学）；Information Technology University（信息技术大学）
We propose a novel deep training algorithm for joint representation of audio and visual information which consists of a single stream network (SSNet) coupled with a novel loss function to learn a shared deep latent space representation of multimodal information. The proposed framework characterizes the shared latent space by leveraging the class centers which helps to eliminate the need for pairwise or triplet supervision. We quantitatively and qualitatively evaluate the proposed approach on VoxCeleb, a benchmarks audio-visual dataset on a multitude of tasks including cross-modal verification,…
Aalto University（阿尔托大学）；University of Amsterdam（阿姆斯特丹大学）；Centrum Wiskunde & Informatica（数学与计算机科学研究中心）；Bristol Robotics Laboratory（布里斯托尔机器人实验室）
Extracting and binding salient information from different sensory modalities to determine common features in the environment is a significant challenge in robotics. Here we present MuPNet (Multi-modal Predictive Coding Network), a biologically plausible network architecture for extracting joint latent features from visuo-tactile sensory data gathered from a biomimetic mobile robot. In this study we evaluate MuPNet applied to place recognition as a simulated biomimetic robot platform explores visually aliased environments. The F1 scores demonstrate that its performance over prior hand-crafted s…
