The paper investigates how dataset imbalance increases vulnerability to backdoor attacks and degrades existing defenses. It introduces Randomized Probability Perturbation (RPP), a black‑box framework that detects poisoned samples using only model output probabilities, providing provable detectability guarantees and a bound on false positives. Experiments on five benchmarks and multiple attacks show RPP outperforms state‑of‑the‑art defenses, especially under class imbalance.
By Miao Lin, Feng Yu, Rui Ning, Lusi Li, Jiawei Chen, Qian Lou, Mengxin Zheng, Chunsheng Xin, Hongyi Wu
The paper investigates whether modern video‑language models can reliably understand temporal dynamics in driving scenarios. Using controlled ego‑motion tasks, the authors find that various prompting strategies are inconsistent and heavily biased, while newer models only show modest few‑shot improvements. However, lightweight LoRA fine‑tuning and probes on intermediate visual features can recover strong motion discrimination, suggesting that motion cues exist within the models but are not effectively leveraged by the final video‑to‑language pathway.
By Shiva Sreeram, Erfan Aasi, Tsun-Hsuan Wang, Alaa Maalouf, Daniela Rus
The paper introduces a vertical federated learning framework that learns one‑step‑ahead cross‑client temporal interdependencies in decentralized industrial systems with heterogeneous, proprietary models. Each client compresses high‑dimensional sensor data into low‑dimensional latent states via a nonlinear state‑space model, while a central Graph Attention Network (GAT) learns a state‑transition model and derives attention coefficients that reflect cross‑client interactions. The authors analytically link these attention coefficients to the Jacobian of the transition model, proving bounded closeness to a centralized oracle, and validate the approach with synthetic and real‑world experiments showing improved federated root‑cause analysis.
By Ayse Tursucular, Ayush Mohanty, Nazal Mohamed, Nagi Gebraeel
The paper presents a radiomics-based approach for classifying thyroid nodules using ultrasound images. It extracts quantitative features with PyRadiomics and employs a Generalized Additive Model (GAM) chosen for its balance of performance and interpretability, achieving an AUC of 0.89, sensitivity of 0.83, and specificity of 0.82 on a public dataset. To provide uncertainty estimates, Cross-Conformal Prediction with label-conditional guarantees is applied, allowing 63% of test nodules to receive a confident single-label prediction while the remaining 37% are referred to clinicians.
By Laura Testa, Pierpaolo Brutti, Umberto Ferraro Petrillo
RARA (Retrieval‑Anchored Representation Alignment) is a two‑stage method for fine‑tuning pre‑trained robot policies on new tasks with few demonstrations. It first retrieves and aligns representations from pre‑training that match the target task’s intent and motion, then fine‑tunes while constraining changes to those retrieved representations. Experiments on Diffusion Policy and GR00T across 13 tasks in simulation and on a real robot show RARA improves in‑distribution success by 9.5 points and out‑of‑distribution success by 13.6 points, outperforming all baselines.
By Xuhui Kang, Sung-Wook Lee, Yan Zhuang, Tianle Zhong, Yen-Ling Kuo
The paper presents a lightweight vocal intensity conversion system that uses a conditional cycle‑consistent GAN to transform speech features so that the output matches any target sound pressure level (SPL) on a continuous decibel scale, rather than relying on discrete categories or reference utterances. The converter operates on features from a frozen speech encoder and is paired with a pretrained vocoder for resynthesis. Experiments comparing log‑mel spectrograms, WavLM, and SpeechTokenizer representations show that the WavLM‑based model achieves 83% of the target intensity change and also shifts the fundamental frequency, with perceptual tests confirming that perceived intensity rises with the target level on amplitude‑normalized stimuli.
By Quentin Le Tellier, Albert Rilliard, Olivier Perrotin, Marc Evrard
The paper presents an information‑theoretic framework that derives a distortion bound for reconstruction attacks on models trained with DP‑SGD. The bound can be computed before training and applies even when an adversary observes the training trajectory and knows all other training data. It separates the information leaked by training (conditional mutual information) from the inherent uncertainty in the data (source entropy) and uses rate‑distortion theory to produce a lower bound on expected reconstruction error, instantiated in both pixel and LPIPS feature spaces.
By Henry Bourne, Robert F. Allison
The paper introduces a reinforcement‑learning framework for managing bundles and flow control in deep‑space networks, where relay nodes must schedule and drop data under intermittent connectivity and limited storage. It defines a goal‑oriented loss based on excess age to prioritize mission‑relevant information and formulates relay control as a constrained semi‑Markov decision process solved with masked Lagrangian PPO. Simulations show that the approach improves age‑based utility and inter‑flow fairness compared to traditional schedulers.
By Sajjad Baghaeea, Aimin Lib, Shanmugarajan Balasubramanian, Elif Uysal
WaSA is a fully unsupervised, offline framework that detects malicious clients in federated learning by having each client send a single privacy‑preserving spectral summary before training. A denoising autoencoder, trained only on clean data, scores clients based on deviation from a learned benign manifold, enabling detection of any attack type without prior labels. Experiments on MNIST, CIFAR‑10, and CIFAR‑100 with up to 90 % malicious clients show consistent improvements over supervised baselines and robust aggregation methods, with negligible communication overhead.
By Alessandro Licciardi
The paper introduces MisVisAgentBench, a paired benchmark with 280 task instances that compare clean versus misleading chart evidence for web agents. It evaluates how deceptive visualizations affect agent execution, providing a diagnostic framework to separate misleading-induced failures from baseline difficulty. Results show that misleading charts reduce task success by over 34 percentage points and tend to redirect agents toward targeted misleading actions.
By Yisheng Li, Dongming Han, Hongkun Pan, Bingcheng Mao, Zijian Wang, Yunlong Mao, Wei Tong, Sheng Zhong
The paper introduces TerraAgent, a training‑free framework that enhances remote sensing visual question answering by integrating three modules: GDR for demonstration retrieval, MACE for evidence‑typed sub‑question decomposition, and CVA for counterfactual evidence auditing. It demonstrates significant improvements on multiple benchmarks, raising accuracy, visual‑dependency scores, and reducing prior‑only answers and hallucinations without fine‑tuning.
By Jingwen Huang, Yiming Fang, Hanxiao Wu, Boyuan He, Keyi Luo
The paper introduces SAMF, a Structure-Aware Matching Framework designed to improve UAV thermal-to-satellite geo‑localization by exploiting stable cross‑modal structural cues. It comprises a Directional Structure Enhancement Module that captures directional and low‑frequency spatial patterns, and a Modality‑Aware Affine Remapping Module that aligns these enhanced features across modalities. Experiments on Boson‑nighttime and Thermal‑UAV‑Paired datasets demonstrate state‑of‑the‑art performance, with ablation studies confirming the complementary benefits of the two modules.
By Chun Liu, Yuxin Wang, Zhenxiang Huo, Linlin Zhu, Tao Xu, Henggang Han
UniMergeTrack introduces a hierarchical model merging strategy for RGB-X visual object tracking, combining independently trained modality-specific trackers into a unified framework. The method merges transferable backbones using high‑pass‑filtered merging, aligns tracking heads with optimal transport‑guided channel alignment, and fine‑tunes with distribution alignment to integrate cross‑task knowledge while preserving modality cues. Experiments on multiple benchmarks show state‑of‑the‑art or competitive performance across RGB‑Thermal, RGB‑Event, and RGB‑Depth tracking tasks, with strong generalization to unseen datasets.
By Shiao Wang, Xiao Wang, Yuhe Ding, Jiarui Xu, Zikang Yan, Bo Jiang, Yonghong Tian, Jin Tang
The paper investigates how different design choices in visual-head discovery scores affect the attention heads selected in vision‑language models. By systematically varying query type, discovery corpus, and attention source in Qwen3‑VL‑8B‑Instruct, the authors find that the attention source and query type have a larger impact on head selection than the corpus. While necessity scores distinguish discovery settings, steerability does not show a significant advantage over random head sets.
By Md Abdul Kadir, Daniel Sonntag
The paper introduces DSG‑GI, a method that adapts pretrained latent video diffusion models for generative inbetweening by using Destination‑Conditioned Adaptation and Semantic‑Geometric Alignment. Destination‑Conditioned Adaptation treats the target frame as the sequence’s endpoint, providing progressive guidance along the generated trajectory. Semantic‑Geometric Alignment aligns token‑wise spatio‑temporal geometry of backbone features with a frozen video encoder and enforces a flow‑matching latent flow constraint near the data endpoint, thereby improving temporal stability, target‑frame faithfulness, and perceptual quality on large‑motion benchmarks.
By Zhiyu Xu (Tianjin University), Wei Shang (Singapore Management University), Yushen Xu (Tianjin University), Wenwen Li (Tianjin Medical University), Zhiyan Xu (Southeast University), Pengfei Zhu (Tianjin University), Dongwei Ren (Tianjin University)
CoVA introduces a lightweight framework that composes VAE features online to provide stage‑wise guidance for flow‑matching training, improving semantic discrimination and spatial structure without extra inference overhead. By jointly training a small latent composer with class supervision, masked reconstruction, and flip consistency, CoVA aligns intermediate SiT features early and later uses detached composer posteriors for sample‑dependent soft‑label conditioning. On ImageNet 256×256, CoVA cuts SiT‑XL/2 FID from 17.2 to 7.4 in 400K steps and achieves an FID of 2.03 in only 400 epochs with classifier‑free guidance, adding just 8.5 M training‑only parameters.
By Zhaowei Chen, Yikai Ge, Linze Li, Yao Tang, Xiang Li
The paper introduces ELMA, a new episodic long‑term motion dataset designed to benchmark anaphoric compositions in text‑to‑motion generation. ELMA contains densely annotated 30‑second clips extracted from in‑the‑wild videos, preserving cross‑segment dependencies that current models struggle to handle. Fine‑tuning state‑of‑the‑art models on ELMA improves both global physical realism and semantic alignment, establishing a robust baseline for long‑term motion synthesis.
By Shun Kato, Dixin Yang, Kotaro Amaya, Yuiga Wada, Kazumi Fukuda, Takashi Shibuya, Yuki Mitsufuji, Mariko Isogawa
SimVTA is a post‑training framework that enhances driving world action models by aligning simulated future videos with candidate trajectories. It samples trajectories, renders high‑reward futures in a real‑world simulator using 3D Gaussian Splatting, and prioritizes video supervision for high‑error futures. The approach improves planning performance, raising PDMS from 86.2 to 91.3 on NAVSIM v1 and achieving strong results on NAVSIM v2 and HUGSIM.
By Changhui Jing, Haochen Tian, Hongchen Li, Tianyu Li, Caojun Wang, Zihao Zhang, Yanhao Liu, Chengen Xie, Bailin Li, Hao Zhao, Zhenjie Yang
The paper presents a new multi-frame representation for event-based action recognition that outperforms single-frame approaches, especially on long and complex event sequences. It employs a masked auto-encoding framework to pre-train a Vision Transformer, using a sparsity-aware masking algorithm, token selection, and multiple mask tokens with position-aware attention to boost performance while managing computational cost. The method achieves state‑of‑the‑art results on the THU‑EACT‑50 CHL and DailyDVS‑200 benchmarks.
By Emma Norrby, Tomas Wilkinson, Per Nyblom, David Gustafsson
The paper introduces CASTER, a framework that formalizes the conversion of synthetic videos into real robot trajectories as a constrained optimization problem. It shows that synthetic videos must probabilistically seed the search within a semantic basin of attraction, allowing local trajectory refinement while maintaining task alignment. By removing simplifying assumptions from existing methods, the authors achieve more robust zero‑shot manipulation and improve task success by 34.1% over the best baseline.
By Yinlong Dai, Homanga Bharadhwaj, Dylan P. Losey