VLGA introduces a dense 3D geometry expert supervised by LiDAR pointmap reconstruction, achieving state-of-the-art safety and driving scores in autonomous driving benchmarks.

Jin Yao, Dhruva Dixith Kurra, Tom Lampo et al.

Proposes DAR-Net, a transformer-based framework with pixel-level scene supervision, achieving 73.33% accuracy in underwater diver activity recognition on the first UDA dataset.

Sadman Sakib Enan, Junaed Sattar

Proposes Turbo-Inference, an iterative inference strategy leveraging detection-segmentation feedback, improving COCO and Cityscapes mAP by over 1% without retraining.

Zhen Zhao, Gang Zhang, Xiaolin Hu et al.

Proposes Bebop with TV loss and rejection sampling to stabilize MTP acceptance rate, achieving up to 95% and 1.8× RL training acceleration.

Yucheng Li, Huiqiang Jiang, Yang Xu et al.

This study compares xLSTM, Mamba-2, and Gated DeltaNet architectures, demonstrating xLSTM's superior performance in complex sequence tasks due to its robust state tracking and memory accumulation.

Anamaria-Roberta Hartl, Levente Zólyomi, David Stap et al.

Proposes Latent World Recovery (LWR), a robust multimodal learning framework that aligns modality-specific embeddings in a shared latent space, handling missing modalities without imputation.

Hui Wang, Tianyu Ren, Joseph Butler et al.

A Unity-based virtual sea trial framework with traceable command-actuation logging supports IMO/ITTC-standard maneuvering data collection for USV system identification.

Paria Rezayan

Fourier features improve point-cloud imitation, raising RoboCasa success from 13% to 34% and real-world normalized score from 14.8% to 40.2%.

Balázs Gyenes, Emiliyan Gospodinov, Jan Frieling et al.

ExtremeWhenBench benchmark reveals that natural-language temporal grounding in long videos is a search problem; a retrieve-then-ground hybrid improves performance by 6.7x.

Sukmin Seo, Geewook Kim

Using Causal Influence Diagrams (CID), this paper proves the impossibility of training AI to honestly report latent knowledge through feedback.

Korbinian Friedl, Francis Rhys Ward, Paul Yushin Rapoport et al.

Proposes a data intermediary model with two-part tariffs to address market failures in AI training data, balancing incentives and model performance.

Yan Dai, Maryam Farboodi, Negin Golrezaei et al.

RankGuard, a decentralized OLTR framework using user click history for robust model filtering, guarantees convergence and resists poisoning attacks.

Marcel Gregoriadis, Martijn de Vos, Sayan Biswas et al.

AGRA enhances robot action accuracy by aligning video features with semantic representations, achieving 80% ID success rate.

Lu Qiu, Yizhuo Li, Yi Chen et al.

SKIM framework compresses procedural knowledge in LLMs, maintaining task performance with 30%-60% compression rate.

Changyue Wang, Weihang Su, Qingyao Ai et al.

This study evaluates decoding-time truthfulness methods (layer-contrast, intervention, logit adapters) on instruction-tuned LLMs, finding limited effects under strict controls, with reasoning prompting more robust.

Ao Sun

Proposes ZOMA, a unified zeroth-order framework combining estimators and acceleration for decentralized minimax optimization.

Haoyuan Cai, Yike Zhao, Aleksandar Armacki et al.

Proposes ISAP-3D with explicit identity-slot alignment for stable 3D part generation, improving structural consistency.

Junlin Hao, Haoshuai Fu, Xibin Song et al.

H-SAL uses self-description text for implicit debiasing, outperforming explicit-label debiasing.

Shun Shao, Zheng Zhao, Anna Korhonen et al.

This paper introduces a metadata-aware multi-prompt reasoning framework for zero-shot accident understanding, achieving a 15% improvement in harmonic mean score on CVPR benchmark.

Tarandeep Singh, Soumyanetra Pal, Soham Biswas et al.

This paper offers a machine learning perspective on uncertainty in dynamical systems, distinguishing aleatoric and epistemic uncertainties, and analyzing their propagation across tasks.

Yusuf Sale, Christopher Bülte, Felix Czaja et al.