CitySTAR introduces a training‑free framework that transforms billion‑scale urban point clouds into a query‑ready scene graph of open‑vocabulary 3D instances, using CodeLLM‑driven tools to supply multimodal evidence for node attributes and spatial relations. It models target‑context topology with paired hypergraphs and performs bidirectional topology verification for structural disambiguation, followed by a Reflective Cross‑modal Grounding module that integrates topology consistency and 2D visual evidence to decide over a metric‑aware 3D context graph. The authors also present CitySTAR‑3D, a benchmark that enhances semantic coverage, instance completeness, bounding‑box fidelity, and spatial‑relation complexity for city‑scale 3D grounding, and report extensive experiments showing consistent improvements in open‑world urban 3D grounding with strong interpretability and generalization.
By Shuai Zhang, Hongye Hou, Qinghe Liu, Zhuoxiao Li, Dongli Wu, Jing Ou, Yuan Liu, Wufan Zhao
The paper introduces the concept of xeno-interpretability, which studies internal distinctions in large language models that lack corresponding human concepts. It distinguishes between human‑interpretable and xeno‑semantic spaces, showing that LLMs possess a far larger internal representational space than can be captured by finite human descriptions. The authors propose an empirical program to identify and characterize these xeno‑representations, noting their potential to influence model behavior in ways that are not fully visible through human‑readable communication.
By F. Pierucci, M. Bracale Syrnikov, M. Prandi, M. Galisai, F. Giarrusso, P. Bisconti
Video DeltaNet (VDN) introduces a hybrid attention mechanism for livestream video generation, combining local Softmax attention with a bidirectional linear memory branch called Video Delta Attention (VDA). VDA updates memory once per frame, integrating spatial tokens, while separate output projections and learnable gates balance the two branches. Applied to MiniMax H3, VDN achieves a 14.5× speedup over the dense baseline, completing 14.3‑second, 768p video denoising in 6.70 seconds on eight NVIDIA B200 GPUs.
By Haocheng Xi, Yiming Xie, Hexu Zhao, Yiwen Zhang, Michael Liu, Thomas Creavin, Kurt Keutzer, Xiuyu Li, Zhaoyang Lv, Chenfeng Xu, Haiwen Feng
The paper argues for a Foundation Model Operating System (FMOS) to virtualize foundation model interactions, similar to how operating systems abstract hardware. Current AI stacks are fragmented, with each framework embedding its own runtime for state, memory, budgets, and guardrails, leading to non-portable behavior and brittle governance. An FMOS would orchestrate knowledge across memory tiers, manage model selection and resource allocation, and enforce verification and policy, learning when to intervene or allow direct inference based on operational experience.
By Suparna Bhattacharya, Tarun Kumar, Cong Xu, Satish Kumar Mopur, Jiahao Li, Ashish Mishra, Aalap Tripathy, Annmary Justine Koomthanam, Martin Foltin, Ian Foster
OceanMoE is a structured conditional sparse Mixture-of-Experts framework designed for long‑horizon multivariate ocean forecasting. It fuses cross‑variable information to build target‑specific local representations and performs content‑conditioned sparse routing at each spatial location, with the number of active experts adjusted by router confidence. Experiments on ORAS5 data show that OceanMoE reduces aggregate forecasting error and maintains lower geometric‑mean normalized RMSE compared to baselines, while expert allocation varies with prediction targets and locations.
By Yishun Zhu, Jian Wang
The paper introduces a digital‑twin framework that simulates opinion dynamics in real Twitter networks by assigning agents attributes such as persona, emotions, centrality, stubbornness, and influence, and using Mistral‑7B to update opinions based on memory and social exposure. Validation on COVID‑19 and U.S. election 2020 datasets shows the framework reproduces opinion trajectories, reducing prediction error by over 50% compared to classical baselines, and improves structural alignment and polarization dynamics. Ablation studies reveal that agent attributes, memory, and social exposure all contribute to predictive fidelity, with agent attributes being the most critical.
By Omran Berjawi, Giuseppe Fenza, Rida Khatoun, Sherali Zeadally
PosteriorBench is a new benchmark that evaluates how well generative inverse solvers recover full posterior distributions rather than just a single reconstruction. It tests four physics-based inverse problems—Darcy flow inversion, Poisson source recovery, carbon capture and storage, and light transport material inference—using high-fidelity reference posteriors generated by rejection sampling and MCMC. The benchmark employs five metrics (posterior-mean error, posterior-standard-deviation error, maximum mean discrepancy, sliced Wasserstein distance, and radially averaged power-spectrum error) to assess pointwise accuracy, uncertainty, distributional alignment, and global frequency fidelity, revealing significant distribution-matching gaps in current solvers and highlighting the importance of neural operators, guidance weights, and generation noise for posterior-variance calibration.
By Jiachen Yao, Zi-Siang Hsu, Xi Deng, Aditi Gupta, Xin Ju, Sally M Benson, Gege Wen, Anima Anandkumar
The paper introduces CalfVO, a monocular visual odometry system that operates without camera intrinsics, test‑time optimization, bundle adjustment, or loop closure. Using a transformer, it predicts relative poses with separate rotation and translation confidences over overlapping image windows, then aggregates these predictions via a confidence‑weighted module to produce a single trajectory. CalfVO achieves the highest accuracy among calibration‑free methods across five benchmarks and runs at 53 FPS, outperforming all baselines.
By Vladimir Yugay, Duy-Kien Nguyen, Theo Gevers, Cees G. M. Snoek, Martin R. Oswald
BioPhys-Bridge is a newly released benchmark dataset designed to evaluate language models on evidence‑grounded scientific reasoning within biophysical literature. Each of its 500 cases includes evidence blocks, stable IDs, quantitative values, units, equations, assumptions, mechanisms, and next‑step decisions, covering six biological domains and nine physical model families. The dataset enforces strict quality gates and has already been evaluated against several models, with DeepSeek‑V4‑Flash achieving the highest evidence‑ID F1 score of 0.360.
By Qingyang Xu
The paper introduces a 202-scenario benchmark to evaluate how large language models (LLMs) handle safety-critical authorization decisions for vehicle voice commands. It tests two local open-weight models and three API-based LLMs, finding alignment scores ranging from 40.1% to 89.1% and noting persistent false execution errors. The study concludes that structured LLM decisions alone are insufficient for safety, recommending an independent enforcement layer to verify tool permissions and vehicle-state constraints before any vehicle function is invoked.
By Diba Afroze, Xingli Zhang, Yazhou Tu, Xiali Hei
FedeRICo is a federated traffic forecasting framework that addresses heterogeneity across client sensor subgraphs by combining gradient-level collaboration with boundary-aware residual communication. It uses a dual-branch architecture: a globally guided branch for transferable forecasting structure and a private residual branch that preserves client-specific corrections and incorporates boundary residual signals. Experiments on four real-world traffic benchmarks show that FedeRICo outperforms state‑of‑the‑art federated spatial‑temporal baselines while keeping training runtime competitive.
By Fermin Orozco, Man Luo, Johan Wahlstr\"om
CoreSense is a robot‑system integration architecture that traces episodic evidence and uses a conflict‑aware belief gate to decide whether to proceed, re‑observe, abstain, or escalates. The gate evaluates scope, provenance, time, contradiction, and support before making a recommendation. Evaluation on public robot datasets, simulations, and a live cloud deployment shows that belief gating can eliminate protocol‑defined unsafe proceeds while maintaining auditability.
By Zoe Li
SCGFM-ART is a structure‑centric graph foundation model that aligns arbitrary graphs onto a shared relational atlas using Amortized Relational Transport (ART). The atlas, defined by a finite set of relational landmarks, provides a universal coordinate system, while ART predicts end‑to‑end graph‑to‑base transport plans, eliminating costly runtime Gromov‑Wasserstein optimizations. The framework decomposes graphs into global relational response coordinates and local node‑to‑role structural correspondences, enabling unified representations that resolve structural and semantic heterogeneity across diverse graph domains.
By Xiaodong He, Xincheng Wang, Zhao Kang
The paper introduces BTS-CAFE, a federated domain generalization framework for respiratory sound classification that addresses stethoscope-induced shortcuts. It combines causality-inspired device-style interventions, counterfactual metadata augmentation, and gradient alignment to reduce style–content entanglement and promote device-invariant decision boundaries. Experiments on ICBHI and SPRSound datasets show a 3.69‑point improvement in out-of-distribution performance over the baseline and outperform conventional data augmentation and federated learning methods.
By Heejoon Koo, Yoon Tae Kim, Miika Toikkanen, June-Woo Kim
CoRe is a model‑agnostic learning objective for direct multivariate time‑series forecasting that replaces pointwise errors with two output‑space constraints: a frequency coherence loss aligning predicted and target spectra, and a low‑rank relational graph loss matching pairwise differences in a PCA subspace. The objective introduces no trainable parameters and can be applied to existing forecasting backbones by changing only the loss. Experiments on standard benchmarks show that CoRe improves strong baselines, compares favorably with recent forecasting objectives, and remains effective across different backbones, datasets, and hyperparameter settings.
By Xiaoyu Lin, Huiran Duan, Yining Liu, Zhixiang Wu, Chu Lin, Lin Lu
CARE‑VI introduces a framework for improving value targets in off‑policy actor‑critic learning by combining Conservative Adaptive Ranking and Screening (CARS), Selector‑Evaluator Value Assessment (SEVA), and Dynamic Adaptive Risk‑aware Enhancement (DARE). CARS limits candidate actions to a budgeted prefix and expands it only when uncertainty exceeds a threshold; SEVA orders candidates with selector critics and reviews their values with an evaluator critic, capping the value at the selector reference; DARE adjusts residual corrections based on candidate reliability and signal gaps. Theoretical analysis bounds errors in each component, and empirical tests on SAC, TD3, and TD7 across four MuJoCo tasks show CARE‑VI consistently outperforms baselines in mean return.
By Xiang Zou, Shengzhu Shi, Junqi Gao, Zhichang Guo
The paper introduces REVEAL, a diagnostic benchmark that stresses Video‑Language Models (VidLMs) on five controlled probes—camera‑motion sensitivity, cross‑frame integration, video sycophancy, language‑only shortcuts, and temporal expectation bias—to assess how well these models encode and use visual evidence. Experiments on 12 VidLMs reveal systematic failures: some visual signals are never reliably encoded, while others are overridden by model priors, leading to performance below chance on several probes that humans solve with high accuracy. Mechanistic probes further pinpoint where and why visual evidence is lost, demonstrating that under assertive prompts a model’s output becomes nearly invariant to real versus random video input, rendering visual evidence causally inert.
By Sethuraman T V, Savya Khosla, Aditi Tiwari, Vidya Ganesh, Rakshana Jayaprakash, Aditya Jain, Vignesh Srinivasakumar, Onkar Kishor Susladkar, Srinidhi Sunkara, Aditya Shanmugham, Rakesh Vaideeswaran, Abbaas Alif Mohamed Nishar, Simon Jenni, Rohan Maheshwari, Derek Hoiem
The paper proposes a Federated Learning framework that integrates an optimized YOLOv8 network for detecting kidney stones in CT images while preserving patient privacy. By enabling multiple medical institutions to collaboratively train a shared model without exchanging patient data, the approach complies with GDPR and HIPAA regulations. Experiments on a distributed CT dataset show a 0.733 mAP@50 and demonstrate fast, real‑time inference suitable for clinical deployment.
By Najiyya Younas, Omar Abdulkader, Yaser Ali Shah, Muhammad Jawad Ikram, Jebran Khan, Amaad Khalil
The paper explores how missing data can inherently enhance privacy in machine learning. By integrating missingness into a differential privacy framework, the authors demonstrate that the absence of certain features can amplify privacy guarantees without altering the underlying algorithm. This reveals a previously overlooked interaction between data incompleteness and formal privacy protections.
By Simon Roburin (LPSM), Rafa{\"e}l Pinot (LPSM), Erwan Scornet (LPSM)
The paper introduces TP‑DATE, a dynamic framework that extends Gromov–Wasserstein optimal transport (GW‑OT) to reconstruct continuous trajectories without simulation. It formulates a broad class of static and dynamic Quadratic‑form OT (QOT) via path actions, proving static‑dynamic equivalence, and develops travelling‑pair flow matching to capture interacting conditional paths in a single vector field. Experiments on synthetic and real spatial transcriptomics data show that TP‑DATE better preserves spatial structure and improves 3D dynamics reconstruction.
By Junda Ying, Zhiwei Zeng, Peijie Zhou, Lei Zhang