The paper introduces Training-Free Omni (TFO), a plug‑and‑play framework that transforms a frozen vision‑language model (VLM) into a speech‑centric omni model without modifying its architecture or requiring multimodal re‑alignment. TFO leverages Whisper to generate confidence‑filtered, timestamped transcripts and routes them through the VLM’s existing language interface, leaving the visual pathway untouched. Evaluations on 56 benchmarks across 21 languages show that TFO matches or surpasses native omni models on audio‑visual tasks, improves audio‑only performance, and preserves strong visual and reasoning capabilities.
By Ankan Deria, Hanoona Rasheed, Xilin He, Fahad Shahbaz Khan, Salman Khan
MultihopSpatial is a new benchmark for Vision‑Language Models that focuses on multi‑hop, compositional spatial reasoning with queries ranging from 1 to 3 hops across varied spatial perspectives. It introduces the Acc@50IoU metric, which jointly evaluates answer selection and precise bounding‑box prediction, and provides a large‑scale training corpus, MultihopSpatial‑Train, to improve spatial intelligence. Evaluation of 37 state‑of‑the‑art VLMs shows that compositional spatial reasoning remains a significant challenge, and reinforcement learning fine‑tuning on the corpus boosts both intrinsic spatial reasoning and downstream embodied manipulation performance.
By Youngwan Lee, Soojin Jang, Yoorhim Cho, Seunghwan Lee, Yong-Ju Lee, Sung Ju Hwang
RoboSPA is a large-scale robotic manipulation dataset and benchmark designed to evaluate Vision‑Language‑Action models on fine‑grained spatial reasoning and long‑horizon procedural planning. It contains 10 task categories, 56 base tasks, and 280 variants across five difficulty levels, with 527K trajectories collected from multiple embodiments and scenes. The benchmark introduces diagnostic metrics beyond binary success, revealing that current VLA models struggle with complex spatial relations, precise execution, and memory‑intensive planning.
By Zhenxuan Fan, Bo Zhang, Yutong Lin, Yuqian Yuan, Juekai Lin, Liang Liang, Zhuoyi Huang, Wenqiao Zhang, Juncheng Li, Siliang Tang, Jun Xiao, Yueting Zhuang
The study investigates how unified vision‑language models (VLMs) can simultaneously support visual understanding and generation. Using controlled benchmarks (SmartWatch and modified CelebA) that pair VQA, captioning, and text‑to‑image tasks, the authors evaluate several LLM‑based architectures built on SigLIP and VQ‑VAE visual spaces. Results show that mixed training can improve both understanding and generation, but the gains depend on how well the visual input and output spaces are aligned; misaligned or distorted visual spaces can weaken or reverse these benefits. The paper also demonstrates that balancing data across tasks and controlling attribute frequencies can help recover underrepresented visual concepts, and that the transfer is driven more by the base language model’s learned relationships than by visual adapters.
By Jihai Zhang, Tianle Li, Linjie Li, Zhengyuan Yang, Yu Cheng
CT‑ΔBench is a new benchmark designed to evaluate vision‑language models on longitudinal 3D medical imaging difference reporting. It provides patient‑level split data, change‑aware metrics, and physician‑validated references to assess clinically meaningful interval changes between two CT scans. The paper also introduces DeltaMed, a baseline model that directly reasons over paired CT scans, and compares it to an indirect two‑stage approach that first generates single‑timepoint reports before differencing.
By Kegeng Tang, Jingbo Wang, Shaogang Ren, Zihao Wang
The article outlines the emerging field of Magnetoencephalography (MEG) foundation models, explaining how these reusable, pretrained models can surpass traditional task‑specific decoding pipelines. It reviews current design choices—such as tokenization, sensor versus source representations, and self‑supervised objectives—and notes the limited number of existing MEG‑specific models and datasets. The authors propose a roadmap that includes native MEG pretraining, adaptation of EEG models, transfer from generic time‑series models, and multimodal integration with other neuroimaging and behavioral data, while emphasizing the need for coordinated infrastructure, rigorous evaluation, and responsible data‑sharing practices.
By Philipp Th\"olke, Hamza Abdelhedi, Yorguin Mantilla-Ramos, Fouad Lbakali, Oumayma Gharbi, Catherine Duclos, Annalisa Pascarella, Vanessa Hadid, Oiwi Parker Jones, Karim Jerbi
WeakMCN introduces a multi-task collaborative network that jointly learns weakly supervised referring expression comprehension (WREC) and segmentation (WRES) using a dual-branch architecture. The WREC branch employs anchor-based contrastive learning and serves as a teacher for the WRES branch, while two novel modules—Dynamic Visual Feature Enhancement (DVFE) and Collaborative Consistency Module (CCM)—facilitate cross-task collaboration. Experiments on RefCOCO, RefCOCO+, and RefCOCOg show significant performance gains over single-task baselines, with up to 3.91% and 13.11% improvements on WREC and WRES respectively, and strong generalization in semi-supervised settings.
By Silin Cheng, Yang Liu, Xinwei He, Sebastien Ourselin, Lei Tan, Gen Luo
The paper introduces Linguistic Trajectory Encoding (LTE), a hybrid representation that compresses dynamic object motion histories using natural language descriptions, sparse spatial anchors, and visual anchors. LTE adapts compression to motion complexity by anchoring periods without reliable observations to the last seen location while preserving accuracy with geometric waypoints and linguistic descriptions. Evaluated on the newly constructed Spatial Memory Benchmark (SMB) from EgoLife multi‑day recordings, LTE achieves 45.3 % success in semantic trajectory retrieval and 48.7 % in long‑horizon object retrieval, outperforming prior structured‑memory and VLM baselines, and compresses trajectories 8.7×–26.1× with sub‑second query latency on 24‑hour video.
By Tianyidan Xie, Shenyi Wang, Qiang Tang, Mingjie Wang, Zhicheng Qiu, Xuanfu Li, Zhan Xu, Jian Yang, Lanjun Wang, Zili Yi
GLASS is a graph‑level anomaly detection framework that aligns graph and language representations on a unit hypersphere to achieve cross‑domain transferability. It constructs a Graph Descriptor Prompt to encode local, global, and semantic graph properties, and uses a multi‑slice soft cosine objective to unify graph and text embeddings. Anomaly scoring is performed via spherical density estimation with von Mises‑Fisher kernels, enabling zero‑shot detection and few‑shot adaptation across twelve benchmarks and three meta‑domains, outperforming recent GLAD baselines.
By Xudong Wang, Chris Ding, Tongxin Li, Jicong Fan
The paper introduces the problem of teaching multimodal large language models (MLLMs) to prioritize must‑have requirements over nice‑to‑have ones in autonomous agent tasks. It evaluates current MLLMs on 3,649 realistic service‑scenario problems and finds widespread failures, then proposes First Things First Reinforcement Learning (FTF‑rl) to explicitly optimize reasoning over multi‑priority user requirements, achieving significant performance gains and generalizing to logical and mathematical reasoning benchmarks.
By Tianjie Ju, Xinyue Xu, Wanxuan Sun, Lingxiao Diao, Gongshen Liu, Zhuosheng Zhang, Cheng Yang
ICM-Bench is a new benchmark for evaluating identity-centric reasoning in multimodal agents with long-term memory. It consists of 839 synthetic video clips totaling 141 minutes and 1,217 open-ended questions about six recurring adults in a one-year life album. The benchmark isolates the ability to maintain recurring person identities and reason over their cross-time relations, and compares various baseline systems, showing that while Gemini 3.1 Pro performs well overall, its accuracy drops on questions requiring long-term identity profiles.
By Shidu Ren, Yunze Liu, Xing Liu, Chi-Hao Wu, Enmin Zhou, Junxiao Shen
The paper introduces LatentStream, a progressive latent working memory framework for streaming video understanding that replaces the traditional store‑and‑retrieve paradigm with a retrieve‑and‑internalize approach. It organizes visual history into short, mid, and long‑term levels using Jenks‑guided adaptive consolidation, then expands memory receptive fields to iteratively retrieve and internalize evidence into a compact latent memory. A confidence‑guided optimization further refines this memory, leading to state‑of‑the‑art performance on online and offline video benchmarks.
By Hongyu Qu, Guangming Yao, Ling Xing, Xiaobin Hu, Rongxing Ding, Guibin Zhang, Fan Zhang, Yi Yuan, Xiangbo Shu, Shuicheng Yan
X-Translator is a low‑cost, modular real‑time speech‑to‑speech translation system that integrates streaming ASR, machine translation, and prompt‑conditioned TTS, managed by a session‑level runtime controller. It uses incremental segment commitment to stabilize ASR streams and an online speaker prompt manager to maintain speaker consistency across multi‑speaker conversations. The system is evaluated on translation quality, speech naturalness, latency, and speaker preservation using OpenSTBench, and its code and demo are publicly available on GitHub.
By Yuxiang Zhao, Yichi Zhang, Yanjie An, Yanqiao Zhu, Zhanxun Liu, Yushen Chen, Qixi Zheng, Haina Zhu, Yunchong Xiao, Keqi Deng, Shuai Fan, Kai Yu, Xie Chen
The paper evaluates uncertainty estimation (UE) methods for clinical vision‑language models (VLMs) on visual question answering (VQA). Across 8 UE techniques and 12 VLMs, UE quality tracks model accuracy, degrading where performance is weakest, and fails to signal uncertainty when models are stressed by hiding the correct answer (NOTA perturbations). However, UE on unperturbed inputs reliably predicts which predictions will collapse under NOTA, suggesting UE can diagnose model fragility.
By Arnisa Fazla, Alberto Testoni, Ameen Abu-Hanna, Barbara Plank, Iacer Calixto
M3T introduces a discrete multi‑modal motion token system for sign language production, addressing the need for non‑manual features such as mouthings, eyebrow raises, gaze, and head movements. The approach couples FLAME’s expressive facial space with SMPL‑X body parameters and uses modality‑specific Finite Scalar Quantization VAEs to achieve high face codebook utilization (99.0%). Trained with an autoregressive transformer and a sign‑to‑text translation objective, M3T outperforms existing methods on three standard datasets, notably improving accuracy on NMFs‑CSL from 49.0% to 58.3% without large‑scale pre‑training.
By Alexandre Symeonidis-Herzig, Jianhe Low, Ozge Mercanoglu Sincan, Richard Bowden
SurgAtlas is the largest surgical video‑language dataset, containing 15,291 videos (2,391 hours) across 18 specialties and over 5,000 procedure types, all sourced from public YouTube. It uniquely includes open‑surgery videos at scale (6,182) alongside more than 9,000 minimally invasive recordings, and introduces standardized benchmarks for open‑surgery video understanding. The dataset offers a rich, multi‑tier annotation schema—segment‑level captions, step/phase descriptions, video‑level surgical narratives, and reasoning‑oriented VQA pairs—validated by experts and built through an automated LLM‑enriched pipeline.
"whyItMatters":"SurgAtlas provides an unprecedentedly large, diverse, and clinically validated resource that can train and benchmark multimodal surgical AI models, advancing the development of next‑generation foundation models for surgery."
By Filippos Bellos, Andre S. Gala-Garza, Miaowei Wang, Alyssa M. Hardin, Ahmad M. Hider, Li Yayuan, Jing Bi, Susan Liang, Chenliang Xu, Donald S. Likosky, Jason J. Corso
InSituMeasure is a new benchmark that tests multimodal large language models on situated measurement tasks in industrial scenes. It includes 2,922 real monitoring images of eight types of engineering instruments, with detailed gauge-attribute labels and noise tags for failure diagnosis. The benchmark defines metrics for numerical accuracy, unit consistency, rejection of unanswerable queries, and alignment of model failures with annotated error factors, revealing that even the best models achieve only 25.7% joint value‑unit accuracy and 51.8% confidence‑diagnosis F1.
By Chao Shen, Xinyuan Li, Yunfan Zhou, Jianguo Yao, Haibing Guan, Zhihai Wang, Xijun Li
WorldReward introduces a vision‑language model–based reward system for camera‑conditioned world models, combining action consistency and visual quality evaluation. It processes paired videos by splitting them into action‑aligned chunks, structuring visual evidence, and aggregating decisions through voting. The model is trained on a large, reasoning‑augmented preference dataset and outperforms GPT‑5.5 on a human‑annotated benchmark, improving both action execution and visual quality when applied to RL post‑training.
By Yibin Wang, Zehan Wang, Junshu Tang, Zhimin Li, Yujie Zhou, Jiazi Bu, Pengyang Ling, Feng Han, Zhixiong Zhang, Long Xing, Shengyuan Ding, Ziang Li, Cheng Jin, Yuhang Zang, Jiaqi Wang, Tianyu Pang
The paper introduces FlexibleFusion, a method for infrared-visible object detection that adapts to both complete and missing-modality scenarios. It employs a Modality-Aware Experts Collaboration mechanism to selectively fuse cross-modal or intra-modal pathways, and a Residual Self-Paced Entropic Optimal Transport module to align heterogeneous feature distributions without heavy optimization. Experiments demonstrate consistent performance across various modality configurations.
By Yue Zhao, Hua Yu, Yukun Zhao, Yuzhi Zhang, Maoguo Gong, Xin Mei, Zhuping Hu, Yanchi Li, A. K. Qin
KnowVis is a framework that converts linear video lectures into knowledge‑centric visual narratives. It first extracts a detailed concept map from multimodal video content to identify key and challenging concepts, then builds structured knowledge units and synthesizes engaging visual summaries. The authors also provide a curated dataset of 125 educational videos across 10 disciplines, paired with 1,079 visual summaries, and show through automated evaluations and a human study that KnowVis produces more accurate, clear visuals that reduce cognitive load and improve learning effectiveness and knowledge retention.
By Yi Xu, Yifan Hou, Xiaoyu Zhang