The paper introduces a new paradigm called VLM-as-probabilistic-grounder, which models the uncertainty of Vision‑Language Model (VLM) predicate groundings as a probability distribution over symbolic states. This probabilistic grounding allows belief‑space planning, producing more robust plans in partially observable settings. Experiments in simulated household robot environments demonstrate that this approach improves robustness and task success compared to deterministic grounding methods.
By Guy Azran, Michael Navat, Sarah Keren
Probe‑VAD introduces an ordinal binary‑probing framework that leverages frozen vision‑language models for training‑free video anomaly detection. By querying ten ordered severity thresholds and extracting YES/NO continuation likelihoods, it builds a cumulative severity profile that is converted into a continuous anomaly score with isotonic projection for ordinal consistency. Experiments on public benchmarks show that this simple interface yields superior performance at low computational cost, avoiding the limitations of caption‑based compression or restricted numerical scoring.
By Jiawei Gu, Qilin Zhao, Tengkuo Guo, Zhiming Zhong, Shuangqing Zhang, Fan Lyu, Fang Zhao, Guo-Sen Xie, Caifeng Shan
arXiv:2604.15678v2 Announce Type: replace
Abstract: Pretrained Vision-Language Models (VLMs) like CLIP show promise in continual learning, but existing Few-Shot Class-Incremental Learning (FSCIL) met...
By Eunju Lee, MiHyeon Kim, JuneHyoung Kwon, Yoonji Lee, JiHyun Kim, Soojin Jang, YoungBin Kim
arXiv:2604.11082v2 Announce Type: replace
Abstract: Visual glitches in video games degrade player experience and perceived quality, yet manual quality assurance cannot keep pace with the growing test...
By Yakun Yu, Ashley Wiens, Adri\'an Barahona-R\'ios, Benedict Wilkins, Saman Zadtootaghaj, Nabajeet Barman, Cor-Paul Bezemer
arXiv:2609.16055v1 Announce Type: cross
Abstract: Test-time compute has emerged as a major approach to improving the capabilities of Large Language Models (LLMs). However, existing test-time reasonin...
By Zhiren Gong, Yikun Hou, Zihao Zeng, Ming Xiao, Chau Yuen, Wei Yang Bryan Lim
The paper reviews ethical and privacy risks of large language model (LLM)–enabled geospatial artificial intelligence (GeoAI), identifying eight recurring issues such as data provenance, spatial privacy, algorithmic bias, and technical risks. It evaluates current responses, noting many remain largely unaddressed or conceptual, and proposes a governance‑aware architecture with enforceable controls illustrated by a flood‑response routing scenario. The authors call for empirical validation, spatially specific interpretability tools, and workforce training to address these emerging risks.
By Maya Subramanian, Devika Jain
The paper introduces MASA, a method for Wild Test-Time Adaptation that uses a frozen multimodal large language model to provide structured semantic anchors, thereby avoiding the self-referential loop common in existing WTTA techniques. MASA selects a small, diverse set of reliability-ranked anchors, encodes their descriptions, propagates them to nearby test samples, and stores this visual‑semantic information in an online prototype memory. The stored descriptors enable lightweight adaptation of normalization parameters, and MASA is evaluated on the WTTA ImageNet‑C benchmark with ResNet and ViT backbones under limited‑batch, mixed‑domain, and imbalanced‑label‑shift scenarios.
By Zhenbin Wang, Lei Zhang, Lituan Wang, Yan Wang, Zhao Zhang, Wei Huang
The paper proposes universal, tool‑based defenses for large language model agents that use external tools, addressing four types of adversarial attacks: direct and indirect prompt injection, memory poisoning, and backdoor attacks. Two main defenses are introduced: Attacker Tool Filtering, which uses anomaly detection to remove suspicious tools, and Normal Tool Recalling, which restores the agent’s original toolset before planning. The authors also add prompt‑based defenses such as Chain‑of‑Thought prompting and self‑reflection, and demonstrate that these methods dramatically lower attack success rates—often to 0%—across multiple open‑source and proprietary LLMs while maintaining or improving task performance.
By Xiaoyan Li, Yunli Wang
OPD‑Aha is a privileged on‑policy distillation method that improves multimodal reasoning by reconstructing the distillation target from the teacher’s isolated visual preference instead of relying on fragile teacher‑student discrepancies. It suppresses continuations that contradict the image, encouraging students to interrupt flawed reasoning with reflection tokens such as "wait" and "actually." This approach leads to consistent improvements across fine‑grained perception and complex multimodal reasoning benchmarks.
By Chenhao Qiu, Dawei Li, Yechao Zhang, Lei Gong, Zhen Tan
The paper introduces VT-Transformer, a model that predicts answerability scores for Visual Question Answering by treating the task as a regression problem rather than a binary classification. It leverages visual and textual features within a Transformer architecture and demonstrates improved performance and robustness on the VizWiz 2020 dataset compared to existing baselines.
By Tung Le, Huy Tien Nguyen, Le Minh Nguyen
RoleBreak is an open benchmark designed to evaluate long‑horizon role‑playing robustness in spoken dialogue systems. It includes 310 character‑based and user‑centered roles, 6,688 human‑verified dialogue turns, and 11,743 fine‑grained evaluation criteria, with 1,856 turns specifically targeting expressive vocal emotion. The benchmark stresses role consistency, interaction quality, safety, and affect over extended conversations, and the authors evaluated nine system configurations across full‑duplex, omni‑modal, and cascaded ASR–LLM–TTS paradigms.
By Yuqi Wang, Fengyuan Liu, Haochen Luo, Zhiqi Yu, Qi Liu
RegRet is a large multimodal model framework that improves region-level retrieval by adding a Region‑Aware Encoder and a multi‑stage training pipeline featuring localized captioning and regional contrastive learning. It also introduces the REGMB benchmark, containing 225k contrastive pairs across four multimodal retrieval tasks. Experiments show RegRet surpasses strong baselines in zero‑shot settings and gains over 20% improvement on REGMB and public benchmarks while maintaining global retrieval performance.
By Xun Liang, Honghui Yang, Weihang Pan, Ruisi Zhao, Boyuan Pan, Yao Hu, Wenxiao Wang, Binbin Lin, Deng Cai
arXiv:2601.18493v2 Announce Type: replace
Abstract: Vision--language models (VLMs) show promise for disaster-response remote sensing, but existing benchmarks mainly emphasize scene-level or damage-ce...
By Sara Tehrani, Yonghao Xu, Leif Haglund, Amanda Berg, Gulnaz Zhambulova, Michael Felsberg
VOR-Bench is a new benchmark for video object removal that addresses shortcomings in current evaluation methods by providing a dataset with paired edited videos and graffiti masks, a realistic motion-capable paired-video acquisition framework (rMPAF), and a perception-driven scoring model (VOR-MDSM). The dataset includes diverse data from model-generated, tool-rendered, and camera-captured sources, ensuring robust real-world assessment. Experiments show that VOR-Bench’s evaluation results correlate strongly (ρ > 0.9) with human subjective judgments, bridging the gap between traditional metrics and human preference.
By Haonan Huang, Tianrui Qiu, Xianghao Zang, Yinan Du, Zhixiang He, Chi Zhang, Hao Sun, Zhongjiang He, Tianwei Cao, Xuchong Zhang, Hongbin Sun, Kongming Liang, Zhanyu Ma
The paper introduces a multimodal anomaly detection framework that uses cross‑modal reconstruction of heterogeneous time‑series sensor data to detect faults in industrial systems. By learning to reconstruct each modality from the others, the method leverages complementary information across sensing channels without requiring explicit temporal alignment or identical sampling rates. An adaptive test‑time thresholding mechanism further improves robustness to distribution shifts caused by changing operating conditions, as demonstrated by strong fault detection performance in three industrial case studies, especially under out‑of‑distribution regimes.
By Magnus Munk Jensen, Dorte Hammersh{\o}i, Rafa{\l} Wi\'sniewski, Olga Fink
AsyncCouple-Flow introduces a new framework for multi‑modal spatio‑temporal forecasting that tackles three key challenges: differing sampling rates, missing modalities, and autoregressive error accumulation. It employs a Modality‑Aware Token Sparsification module to produce equal‑length sequences, an Asynchronous Cross‑Modal Coupling Graph to fuse data under arbitrary asynchrony and missingness, and a Flow‑Matching Forecasting Head that models multi‑step prediction as a conditional ODE. Experiments on weather and traffic datasets demonstrate that the method outperforms state‑of‑the‑art baselines and remains robust even when up to two modalities are missing.
By Zhixiang Wu, Yining Liu, Bo Zhao, Szu-Yu Chen, Huiran Duan, Chu Lin, Chuanguang Yang
NeuroTS-Net is a 3‑D encoder‑decoder CNN designed for multi‑class semantic segmentation of pediatric brain tumors in multi‑modal MRI. It uses a dual‑scale raw‑detail stream, adaptive low‑resolution context selection, and detail‑preserving multipath downsampling to maintain fine intensity and boundary information while modeling broader tumor context. Trained on the BraTS 2026 pediatric dataset, it outperformed nnU‑Net and MedNeXt, achieving Dice scores of 0.938/0.937 on internal validation and 0.927/0.926 on the official challenge set.
By Darius Peteleaza, Razvan-Gabriel Dumitru, Bogdan Neamtu, Arpad Gellert, Mariana Sandu, Claudiu Matei
Intrinsic Robot Rewarding (IRR) leverages existing vision‑language‑action (VLA) systems to evaluate a robot’s own outcomes and provide feedback for policy improvement. By using successful demonstration endpoints as task‑specific references and the policy’s frozen visual encoder as the feature space, IRR adds a reference bank and scoring operation to the current pipeline without requiring a separate evaluator or additional perception backbone. The approach aims to reduce integration effort, reward computation cost, and human outcome scoring while enabling learning from the data already available in industrial robot systems.
By Tobias Schaffer, Mohab Elkhayat, Daniela Nicklas, Mustafa Almohamad, Elham Al-Fuqara
The paper introduces DiaWhisper-DPO, an end‑to‑end model that fine‑tunes Whisper-large-v3 with LoRA and a frame‑level role head to transcribe and attribute utterances in clinical interviews. It further refines the system using failure‑mined preference optimization (DPO) that leverages genuine decoding failures as rejected completions, eliminating the need for human preference data. On the DAIC‑WOZ dataset, DiaWhisper‑DPO attains 0.973 role accuracy and 0.119 DER, outperforming cascaded baselines by 72% and dramatically reducing seed variation, while also improving performance on the cross‑lingual PDCH‑HAMD dataset.
By Weiming Li, Ana Catarina Fidalgo Barata, Miguel Constante, Jo\~ao Miguel Sanches
The paper examines how the tight integration of multimodal understanding and image generation in Unified Multimodal Models (UMMs) can create new safety vulnerabilities. It introduces RICE, an attack framework that exploits bidirectional interactions between generation and understanding to propagate unsafe signals across modalities. Experiments demonstrate high attack success rates in both Generation‑to‑Understanding and Understanding‑to‑Generation pathways, revealing previously overlooked safety weaknesses in UMMs.
By Kaishen Wang, Heng Huang