Artificial Intelligence
Artificial Intelligence news — Monday 05 October 2026
150 stories from 11 sources, every one in English. Each links to its publisher.
-
40% discount for Google, 'AI Pro' domestic only.. . Bulletin Board
AI Times (Korea)
Translated from Korean
-
Clarification Samsung Hospital, Park, Soo - Jin Bae Yong-ICU visitation privileges, however...
AI Times (Korea)
Translated from Korean
-
Antarpic, ‘Voice Interviews’ Up until October 6th...Combined the Public and Sound Learning Collection
AI Times (Korea)
Translated from Korean
-
New Links, 50,000 Hours Brain Data Pre-Learning...BCI also opened the ‘Powered Model’ era
AI Times (Korea)
Translated from Korean
-
‘Human vs Robot’ Cage Fight, California’s ‘Dual Command’ received
AI Times (Korea)
Translated from Korean
-
KAIST develops ‘CONDA’ technology that keeps scanning accuracy even when data changes
AI Times (Korea)
Translated from Korean
-
a16z "U.S. furniture 2.2% only AI pay subscription...technical led circulation core"
AI Times (Korea)
Translated from Korean
-
GPT-6 Astra, 217 Napoleon Password Decryption in 6 hours
AI Times (Korea)
Translated from Korean
-
MIT·Sakana, self-extinguishing AI agent ‘SIFT’ announced.. .'Resolution of the illusion cost bottleneck'
AI Times (Korea)
Translated from Korean
-
Amazon, the 11th largest AI chip investor, after sale, Lis...
AI Times (Korea)
Translated from Korean
-
... Google, Open Source Bug Bounty Suspended
AI Times (Korea)
Translated from Korean
-
ChatGPT clothes and shopping... OpenAI Introduces 'Watch Wear'
AI Times (Korea)
Translated from Korean
-
Connecting AI agents to enterprise knowledge
MIT Technology Review Biotech
For all the data that AI systems continually amass and analyze, enterprise AI agents often suffer from a curious shortcoming: a lack of knowledge. More than data, knowledge is the understanding of what the data means in the context of…
-
Broadcom guarantees 80 trillion loans in Antwerp... Financial support for chip sales
AI Times (Korea)
Translated from Korean
-
"Present American Criminal Libraries"...China Hacker, AI Policy Expert Account
AI Times (Korea)
Translated from Korean
-
Netherlands large eyewear chain, close-up to meta smart glasses sales
AI Times (Korea)
Translated from Korean
-
Our approach to EU text provenance rules
OpenAI Blog
How OpenAI is approaching text watermarking under EU rules. Learn where watermarks apply, how detection works, and why access starts with researchers.
-
Computer Vision: SIFT algorithm (Scale Invariant Feature Transform)
Towards Data Science
Elegant object matching from several viewpoints The post Computer Vision: SIFT algorithm (Scale Invariant Feature Transform) appeared first on Towards Data Science .
-
Physical AI moves beyond traditional robotics
AI Business
Physical AI is moving beyond traditional robotics into real enterprise tasks, but enterprises still face big challenges integrating and scaling robotic systems.
-
Bringing predictive analytics to the agentic AI era
MIT Technology Review Biotech
In 2026, the question for enterprise AI is no longer whether predictive models can outperform statistical forecasts—that argument is settled. The big question now is how to enable predictive systems to act on their own conclusions without…
-
Water Isotopes Offer A New Way To Improve Weather Forecasts
Asian Scientist
A study using satellite observations from 2014 to 2019 finds that adding water isotope data to weather models can improve forecasts by up to five days.
-
Import AI 475: Swarm scaling; Google DeepMind watermarks biology; and the AI science economy
Import AI Newsletter
Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now When should you use swarms? When you are in a hurry: …How…
-
The Download: AI’s popularity paradox and EmTech Future 2026
MIT Technology Review Biotech
This is today’s edition of The Download , our weekday newsletter that provides a daily dose of what’s going on in the world of technology. People really hate AI, so why can’t they get enough? —Will Douglas Heaven Over the summer I talked…
-
"Create the code of the dislike"...Astra, StarCraft Benchmarker
AI Times (Korea)
Translated from Korean
-
OpenAI 'Agent Blast'... The world’s legal and regulatory warning storm
AI Times (Korea)
Translated from Korean
-
How to Build a Cheap, Yet Reliable Model Router With Jev
Towards Data Science
Model routing can finally be a design choice rather than a system enhancement The post How to Build a Cheap, Yet Reliable Model Router With Jev appeared first on Towards Data Science .
-
"Perfect, in the wrong hand, extreme risk"...AI safe warning
AI Times (Korea)
Translated from Korean
-
Building advertising for the way people use AI
OpenAI Blog
OpenAI introduces a new visual ad format in ChatGPT and expands measurement tools, attribution partnerships, and brand suitability for advertisers.
-
"AI Victim Video" in the middle.. . The U.S. Court "Morning was wrong"
AI Times (Korea)
Translated from Korean
-
Expanded to 1MW in Australia and solar support...Medical equipment is applied from October
AI Times (Korea)
Translated from Korean
-
Opened the 9th day of the Suncheon National Fisheries...
AI Times (Korea)
Translated from Korean
-
Goalkeeper Kim Byung-ji pony tail nerve damage in a traffic accident legend
AI Times (Korea)
Translated from Korean
-
China Researcher, Bamboo Water Solar Structural Test...Bolt thickness left
AI Times (Korea)
Translated from Korean
-
Jeonnam-Gwangju, bus terminals, elderly and seniors
AI Times (Korea)
Translated from Korean
-
“AI warns mankind” warns Jacob Coxon, New York City Council
AI Times (Korea)
Translated from Korean
-
Design AI Loop
AI Times (Korea)
Translated from Korean
-
MSK, 'SpaceXSI', a branding s...Trumf 'SI'
AI Times (Korea)
Translated from Korean
-
Trump Launches ‘Super Intelligence Bag’...National Information Bureau, White House AI Carre
AI Times (Korea)
Translated from Korean
-
People really hate AI, so why can’t they get enough?
MIT Technology Review Biotech
Over the summer I talked to the CEO of Springboards, a startup building an LLM that’s designed to come up with a wider variety of responses than its mainstream rivals do. At the start of the call, he said something that’s been stuck in my…
-
LPA-CWM: A Learned Physical Adjudicator for Motion Reasoning with Counterfactual World Models
arXiv cs.CV
arXiv:2609.14073v2 Announce Type: replace Abstract: Counterfactual world models (CWM) extract motion from pretrained video predictors by comparing factual and intervened predictions, but uniform aggregation weights responses equally…
-
PatchScene: Patch-based Voxel Diffusion for Large-Scale Scene Completion
arXiv cs.CV
arXiv:2606.03915v2 Announce Type: replace Abstract: We propose PatchScene, a novel diffusion-based framework for large-scale LiDAR scene completion. Unlike existing methods that rely on global latent representations or dense voxel grids…
-
Does Every User Need a Private LoRA? Decoupling Personalization from Per-User Adaptation
arXiv cs.CL
arXiv:2610.02353v1 Announce Type: cross Abstract: Personalized large language models often require a complete adaptation state for each user. However, this paradigm scales poorly as the user population grows. We revisit this design through…
-
Understanding Affective Adaptation in Multimodal Foundation Models: Emergent Functional Specialization
arXiv cs.CV
arXiv:2601.15906v2 Announce Type: replace Abstract: Despite rapid progress in multimodal affective foundation models, how affective capabilities emerge within their internal architectures remains poorly understood. A critical open question…
-
Adaptive Second-Order Solvers for Fast Stochastic Diffusion Sampling
arXiv cs.CV
arXiv:2610.03034v1 Announce Type: cross Abstract: Diffusion models rely on numerical solvers requiring time-discretization, which has a large influence on the tradeoff between sampling cost and quality. However, the computational…
-
A Manifold-Aware Topic Modeling Approach via Rank-Based Prototypes
arXiv cs.CL
arXiv:2609.29630v1 Announce Type: cross Abstract: Recent topic models leverage pretrained embeddings, but neural architectures produce latent representations without grounding in specific texts, and clustering-based pipelines assign…
-
Learning the Latent Structure: A Feature-Centric Approach to Graph Data Augmentation
arXiv cs.LG
arXiv:2610.02517v1 Announce Type: new Abstract: Graph-structured data plays a pivotal role in modeling complex relationships. However, real-world graphs are often incomplete due to data collection and observational constraints, severely…
-
ProAR: Learning Prospective Reasoning with Autoregressive Video Models
arXiv cs.CV
arXiv:2610.03664v1 Announce Type: new Abstract: Autoregressive (AR) video models excel at causal generation, but their reliance on next-chunk prediction confines them to a short-sighted, reactive paradigm. This limitation is particularly…
-
FRUC: Feedforward Dynamic Scene Reconstruction from Uncalibrated Collaborative Driving Views
arXiv cs.CV
arXiv:2605.29997v2 Announce Type: replace Abstract: We present FRUC, a feedforward 3D Gaussian Splatting framework for dynamic scene reconstruction from uncalibrated collaborative driving views. Existing multi-agent reconstruction…
-
A Vision-Language Model (VLM)-based Pipeline for End-to-End Procedural Modeling of Field-Grown Maize from Point Clouds
arXiv cs.CV
arXiv:2610.03468v1 Announce Type: new Abstract: Editable 3D models of field-grown crops support high-throughput phenotyping and in silico breeding trials, but building them from scanned point clouds requires organ-level segmentation and…
-
Single-Pass Uncertainty Heads for Claim-Level Hallucination Detection in Persian Medical Language Models
arXiv cs.CL
arXiv:2610.03482v1 Announce Type: new Abstract: Hallucination detection is particularly important for medical language models, but repeated-sampling approaches are expensive and existing uncertainty-head resources do not directly transfer…
-
An Elastic Shape Variational Autoencoder for Skeleton Pose Trajectories
arXiv cs.CV
arXiv:2605.09231v4 Announce Type: replace Abstract: Deep generative models provide flexible frameworks for modeling complex, structured data such as images, videos, 3D objects, and texts. However, when applied to sequences of human…
-
GB-LSR: Local Spectral Decoding with a Learned Global Bandwidth for Arbitrary-Scale Super-Resolution
arXiv cs.CV
arXiv:2606.19617v2 Announce Type: replace Abstract: We present GB-LSR (Global-Bandwidth Local Spectral Representation), a fixed-grid local spectral representation for continuous image decoding. The image domain is partitioned into…
-
XClipGS: Exact Half-Space Clipping for Medical Volume Gaussian Splatting
arXiv cs.CV
arXiv:2608.07760v2 Announce Type: replace Abstract: Gaussian-splatting proxies enable interactive rendering of volumetric medical scans, but a clipping plane exposes anatomy not constrained by external-view training and intersects…
-
RetiWave-Mamba: A Dual-Stream Network for Retinal Disease Detection based on Multi-scale Context and Feature-Adaptive Mamba Projection
arXiv cs.CV
arXiv:2608.17623v2 Announce Type: replace Abstract: Retinal diseases are a leading cause of irreversible vision impairment, making early and accurate diagnosis essential for effective treatment. Optical Coherence Tomography (OCT) serves as…
-
Llama-Mobile: Efficient 2.7-Bit Quantization of VLMs
arXiv cs.CV
arXiv:2608.21134v2 Announce Type: replace Abstract: Deploying vision-language models (VLMs) on mobile devices is challenging due to their significant memory and compute requirements. We present a framework for quantizing VLMs for efficient…
-
Embedded Bi-Temporal Building Damage Assessment for On-Board Data Reduction
arXiv cs.CV
arXiv:2609.37013v2 Announce Type: replace Abstract: Rapid assessment of building damage after natural disasters is essential to support emergency response. Earth Observation satellites can acquire relevant imagery shortly after an event…
-
TomoTransformer: Towards a Foundation Model for CT Reconstruction
arXiv cs.CV
arXiv:2609.37605v2 Announce Type: replace Abstract: Supervised deep learning has advanced sparse-view tomographic reconstruction. However, conventional models, which typically map filtered back-projection (FBP) images or sinograms to clean…
-
Color Independent Word Segmentation From Transcribed Bangla Passages
arXiv cs.CV
arXiv:2610.01191v2 Announce Type: replace Abstract: An optical character recognition(OCR) system can scan paper and extract text, making people's jobs easier. While numerous OCR systems are accessible in the software sector, finding a…
-
WakeKV: Reactive, Reversible KV Residency for Heads That Change Their Minds
arXiv cs.CL
arXiv:2610.02713v1 Announce Type: new Abstract: Most KV-cache compression methods classify attention heads once, either offline or during prefill, and keep this classification fixed throughout generation. Across three models (1.5B-8B) and…
-
Do Large Language Models Know Colombian Law? A Reliability Benchmark for the Colombian Legal System
arXiv cs.AI
arXiv:2610.03639v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to support legal practice, education, and research, yet their reliability in national legal systems outside the United States remains…
-
Writerslogic at PAN 2026: Process over Content for Robust Detection under Domain Shift
arXiv cs.CL
arXiv:2610.03565v1 Announce Type: new Abstract: We describe the Writerslogic systems for three PAN at CLEF 2026 shared tasks (Reasoning Trajectory Detection, Voight-Kampff Generative AI Detection, and Multi-Author Writing Style Analysis)…
-
FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training
arXiv cs.CL
arXiv:2610.02828v1 Announce Type: cross Abstract: Adaptive LLM reinforcement-learning post-training changes multiple training actuators online, including rollout temperature, group size, clipping, KL regularization, verifier allocation…
-
Cryo-Bench: Benchmarking Foundation Models for Cryosphere Mapping
arXiv cs.CV
arXiv:2603.01576v5 Announce Type: replace Abstract: Geo-Foundation Models (GFMs) have been evaluated across diverse Earth observation tasks and domains, showing strong potential to produce reliable maps even with sparse labels. However…
-
CHASE-VLA: Post-Training Quantization Framework for Vision-Language-Action Models with Chunk-Aware Scale Estimation
arXiv cs.CV
arXiv:2610.02666v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models map visual observations and language instructions to continuous robot actions, but a diffusion-based action expert (AE) poses a key challenge for low-bit…
-
The Effective Depth Paradox: Topology and Trainability in Deep CNNs
arXiv cs.CV
arXiv:2602.13298v4 Announce Type: replace Abstract: This paper presents a controlled comparative study of convolutional neural network (CNN) topology and image classification performance across the architectural families VGG, ResNet, and…
-
When Is Accuracy Evidence? A Unified Theory of Generalisation, Validation, and Information Fusion
arXiv cs.LG
arXiv:2610.03465v1 Announce Type: cross Abstract: K-fold cross-validation (CV) is widely used as evidence of out-of-sample performance, although folds are neither independent experiments nor equally informative under heterogeneous data…
-
WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites
arXiv cs.CV
arXiv:2610.03036v1 Announce Type: cross Abstract: We present WebFovea, a vision-based web agent that placed 2nd in the WebRetriever Challenge 2026 with a final score of 57.0 out of 100. The challenge evaluates agents end to end on Protocol…
-
EVEWorld: Physical Evolution Supervision for Embodied World Models
arXiv cs.CV
arXiv:2610.03374v1 Announce Type: new Abstract: Embodied world models enable scalable simulation of embodied interactions for robot learning. However, existing models are prone to Model Laziness, as they focus on visual fidelity at the…
-
Not Until the Evidence Says So: Teaching LLM Investigators When to Close a Case
arXiv cs.CL
arXiv:2610.03190v1 Announce Type: new Abstract: Accident, defect and outage investigations end with a decision that ordinary question answering never faces: whether the evidence gathered so far is enough to close the case. We study this…
-
ArrivalBench: Agent-Generated Data Pipelines Are Correct Once and Wrong Under Time
arXiv cs.LG
arXiv:2610.02363v1 Announce Type: new Abstract: Benchmarks for agent-generated data work grade a pipeline by running it once against a fixed snapshot. ArrivalBench instead re-executes the pipeline an agent leaves behind under adversarial…
-
World Action Modeling with Progressive Visual Planning
arXiv cs.CV
arXiv:2610.02508v1 Announce Type: cross Abstract: World action models (WAMs) have emerged as a promising paradigm for robotic control by jointly predicting future visual dynamics and actions from an initial observation and instruction…
-
Corrupted but Correct: Why Vision-Language Models Lie to Themselves Internally
arXiv cs.CV
arXiv:2610.03445v1 Announce Type: new Abstract: A targeted adversarial perturbation can drive a vision-language model's (VLM's) teacher-forced training loss for a fixed target caption to near zero, yet the same model, allowed to generate…
-
CONTRA: Discovering and Qualifying Behavior-Changing Questions for Selective Clarification in LLM Code Generation
arXiv cs.AI
arXiv:2610.01769v1 Announce Type: cross Abstract: Coding agents can generate code that appears correct but implements behavior the user never intended. This mismatch can arise when an agent silently resolves underspecified requirements…
-
VDOT++: Unified Few-Step Video Generation via Unbalanced Optimal Transport Distillation
arXiv cs.CV
arXiv:2610.03221v1 Announce Type: new Abstract: Video creation spans text-to-video (T2V), image-to-video (I2V), and condition-based generation, yet video diffusion models remain costly because they repeatedly evaluate large backbones…
-
Cordial Learning: Distributed Training with Correlated Data
arXiv cs.LG
arXiv:2610.03330v1 Announce Type: new Abstract: We consider a distributed learning task with agents that have correlated data. Specifically, the label of an agent depends on the input of other agents for the same sample, and these inputs…
-
CriticHack: Evaluating Visual Rewards Under Robot Policy Optimization
arXiv cs.AI
arXiv:2610.02527v1 Announce Type: cross Abstract: Learned visual reward models are increasingly used to optimize robot policies, yet a reward model can score an execution that acts on the wrong object as highly as one that completes the…
-
Enhancing Biomedical Named Entity Recognition via Multiple Programming Languages Instruction Tuning and Ensemble Method
arXiv cs.CL
arXiv:2610.02949v1 Announce Type: new Abstract: Instruction tuning has become a common paradigm for applying large language models (LLMs) to biomedical named entity recognition (BioNER). However, existing instruction-tuning approaches…
-
TRAC: Trajectory-aware Reuse and Adaptive Correction for Efficient Autoregressive Video Generation
arXiv cs.CV
arXiv:2610.02779v1 Announce Type: new Abstract: In this paper, we present trajectory-aware reuse and adaptive correction (TRAC), a training-free framework for efficient autoregressive (AR) video generation. Existing acceleration methods…
-
Understanding Trajectory Heterogeneity in Federated World Model Learning
arXiv cs.LG
arXiv:2610.02957v1 Announce Type: new Abstract: World models learn state evolution from trajectories, making access to temporal context a central training requirement. Federated learning can use distributed records, while ownership…
-
DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents
arXiv cs.CV
arXiv:2610.02320v1 Announce Type: new Abstract: Computer-use agents need to reliably ground action targets in complex desktop scenes, where multiple applications, overlapping windows, and visually similar controls compete for attention…
-
On Unlearning for Time-series Forecasting
arXiv cs.LG
arXiv:2610.02865v1 Announce Type: new Abstract: Time-series forecasting is widely used in sensitive domains. Models in these settings are often trained on longitudinal user- or entity-level records, which may later require removal because…
-
A Two-Stage Cascade for Near-Real-Time Forest Anomaly Detection from Sentinel-1 SAR Time Series
arXiv cs.LG
arXiv:2610.02763v1 Announce Type: new Abstract: Tropical forest monitoring is essential for global climate stability and biodiversity preservation. To address the urgent need for rapid, reliable detection of forest loss which is essential…
-
AIGS: Adaptive Incremental Gating System for Online Representation Learning in Non-Stationary Data Streams
arXiv cs.LG
arXiv:2610.02661v1 Announce Type: new Abstract: Real-time data streams in Web of Things (WoT) and edge computing environments often evolve through latent regime changes. For online representation learning under strict computational…
-
WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites
arXiv cs.LG
arXiv:2610.03036v1 Announce Type: new Abstract: We present WebFovea, a vision-based web agent that placed 2nd in the WebRetriever Challenge 2026 with a final score of 57.0 out of 100. The challenge evaluates agents end to end on Protocol…
-
Page-EntroKV: Hardware-Aligned, Entropy-Weighted KV-Cache Eviction under Grouped-Query Attention
arXiv cs.LG
arXiv:2610.03135v1 Announce Type: new Abstract: Serving long-context autoregressive language models is constrained by the key-value (KV) cache. Most dynamic eviction methods score token importance per query head and choose tokens…
-
S$^{2}$-PINN: Stochastic Separable Physics-Informed Neural Networks
arXiv cs.LG
arXiv:2610.03303v1 Announce Type: new Abstract: Uncertainty quantification (UQ) for random partial differential equations (PDEs) is ubiquitous in computational science and engineering. However, classical spectral solvers for this class of…
-
Fisher-Guided Submodular Data Selection for Continual Pre-Training of Large Language Models
arXiv cs.LG
arXiv:2610.02593v1 Announce Type: new Abstract: Data selection is already a central bottleneck in large-language-model training, where web-scale corpora are noisy and token budgets are finite. In continual pre-training (CPT), it becomes a…
-
AIBL: Augmented Instance-Based Learning with Structured Memory and Neural Embeddings
arXiv cs.LG
arXiv:2610.03413v1 Announce Type: new Abstract: Sequential learning systems often make decisions from accumulated experience while receiving high-dimensional inputs whose distribution may change over time. Instance-Based Learning Theory…
-
Get a GRIP, this will be a long TRIP: A Quantifiable Long-Range Framework for Verifying Over-squashing
arXiv cs.LG
arXiv:2610.03556v1 Announce Type: new Abstract: Empirical claims about the connection between over-squashing and long-range interactions in GNNs, can only be trusted if the benchmarks used to validate them genuinely require long-range…
-
DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents
arXiv cs.LG
arXiv:2610.02320v1 Announce Type: cross Abstract: Computer-use agents need to reliably ground action targets in complex desktop scenes, where multiple applications, overlapping windows, and visually similar controls compete for attention…
-
Where Quantum Fourier Sampling Stops Short: A Three-Gate Audit Protocol for Delay-PUF Security Models
arXiv cs.LG
arXiv:2610.02636v1 Announce Type: cross Abstract: Quantum Fourier sampling may help audit the spectral learnability of delay-based physical unclonable functions (PUFs). We ask whether that promise survives access matching, a strong…
-
FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training
arXiv cs.LG
arXiv:2610.02828v1 Announce Type: cross Abstract: Adaptive LLM reinforcement-learning post-training changes multiple training actuators online, including rollout temperature, group size, clipping, KL regularization, verifier allocation…
-
Mixture-of-Experts for Cryptocurrency Order Execution: Training Stability, Tail Risk, and Failure Modes
arXiv cs.LG
arXiv:2610.03369v1 Announce Type: cross Abstract: Deep reinforcement-learning policies for order execution can vary substantially across training seeds, so apparent architectural gains may reflect favourable training realisations rather…
-
From Patching to Pruning Visual Computation in Vision Language Models
arXiv cs.LG
arXiv:2610.03389v1 Announce Type: cross Abstract: Vision language models (VLMs) incur substantial inference cost because every visual token is processed by the attention and MLP projections of every decoder layer, even when token-specific…
-
MGhana-ST: A Low-Resource Speech Translation Dataset for Ghanaian Languages and an Analysis of Multilingual Training Trade-offs
arXiv cs.CL
arXiv:2609.40041v2 Announce Type: replace Abstract: We present MGhana-ST, a speech translation dataset for four low-resource Ghanaian language varieties: Ga, Twi (Akuapem and Asante), Ewe, and Fante. MGhana-ST is an ongoing annotation…
-
Differential Privacy as a Perk: Federated Learning over Multiple-Access Fading Channels with a Multi-Antenna Base Station
arXiv cs.LG
arXiv:2510.23463v4 Announce Type: replace Abstract: Federated Learning (FL) is a distributed learning paradigm that preserves privacy by eliminating the need to exchange raw data during training. In its prototypical edge instantiation with…
-
System-Prompt Anchoring with Cross-Attention Layers
arXiv cs.LG
arXiv:2605.09737v3 Announce Type: replace Abstract: Cross-attention provides a dedicated route from a selected information source into a model's computation, but the effect of where that route is inserted remains underexplored. We study…
-
HINT-SD: Targeted Hindsight Self-Distillation for Long-Horizon Agents
arXiv cs.LG
arXiv:2605.17873v3 Announce Type: replace Abstract: Training long-horizon LLM agents with reinforcement learning is challenging because sparse outcome rewards reveal whether a task succeeds, but not which intermediate actions caused the…
-
Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts
arXiv cs.LG
arXiv:2607.19060v2 Announce Type: replace Abstract: Fast prediction of the response of adhesive soft viscoelastic contacts represents a current challenge in soft robotics and for gripping and manipulation tasks. Determining the complete…
-
Exact Risk Ratios for Weighted Data Selection in Linear Regression
arXiv cs.LG
arXiv:2608.28007v2 Announce Type: replace Abstract: How much data must a fixed learner retain? Hanneke, Moran, Shlimovich and Yehudayoff (COLT 2025) posed this question for linear regression with the minimum-norm empirical risk minimizer…
-
MASCIT: A Mask-Aware State Space Classifier for Naturally Irregular Time Series
arXiv cs.LG
arXiv:2609.34409v2 Announce Type: replace Abstract: Naturally irregular time series combine asynchronous observations, missing values, unequal lengths, and nonuniform sampling, while dense adapters can discard temporal structure. We…
-
The Conflict Between Logic and Memory: Training Conditions for Optimizer-Dependent Rule Acquisition
arXiv cs.LG
arXiv:2610.00403v2 Announce Type: replace Abstract: Optimizers can fit the same task while acquiring different generalizing relations. We study the training conditions governing these differences in single-hidden-layer ReLU networks…
-
Token Space: A Category Theory Framework for AI Computations
arXiv cs.LG
arXiv:2404.11624v4 Announce Type: replace-cross Abstract: Token Space is a categorical framework and mathematical language for AI computations. It connects internal relations, program descriptions, execution states and observations…
-
Asymptotic Performance of Time-Varying Bayesian Optimization
arXiv cs.LG
arXiv:2505.13012v3 Announce Type: replace-cross Abstract: Time-Varying Bayesian Optimization (TVBO) is the go-to framework for optimizing a time-varying black-box objective function that may be noisy and expensive to evaluate, but its…
-
Bifidelity Karhunen-Lo\`eve Expansion Surrogate with Active Learning for Random Fields
arXiv cs.LG
arXiv:2511.03756v2 Announce Type: replace-cross Abstract: We present a bifidelity Karhunen--Lo\`{e}ve expansion (KLE) surrogate model for field-valued quantities of interest (QoIs) under uncertain inputs. The QoIs considered here are…
-
Power-SMC: Low-Latency Sequence-Level Power Sampling for Training-Free LLM Reasoning
arXiv cs.LG
arXiv:2602.10273v3 Announce Type: replace-cross Abstract: Reasoning ability in large language models is often attributed to \emph{distribution sharpening}: concentrating output probability on high-likelihood sequences. Recent works show…
-
Uncertainty Quantification for Flow-Based Generalist Robot Policies
arXiv cs.LG
arXiv:2606.18043v2 Announce Type: replace-cross Abstract: Generalist robot policies, such as vision-language-action models (VLAs) and world-action models (WAMs), combine powerful pretrained backbones with expressive generative action heads…
-
GB-LSR: Local Spectral Decoding with a Learned Global Bandwidth for Arbitrary-Scale Super-Resolution
arXiv cs.LG
arXiv:2606.19617v2 Announce Type: replace-cross Abstract: We present GB-LSR (Global-Bandwidth Local Spectral Representation), a fixed-grid local spectral representation for continuous image decoding. The image domain is partitioned into…
-
ORACLE: Agentic AI Orchestrator Routing Via Adaptive Verifier Calibration Feedback
arXiv cs.LG
arXiv:2607.22465v4 Announce Type: replace-cross Abstract: Modern enterprise agent deployments consist of a heterogeneous pool of large language models (LLMs) having diverse capabilities and cost. Existing model routing strategies optimize…
-
AECSF: Adaptive Ensemble Conditional Score Filtering for High-Dimensional Nonlinear Data Assimilation
arXiv cs.LG
arXiv:2609.32411v2 Announce Type: replace-cross Abstract: Bayesian state estimation for high-dimensional nonlinear dynamical systems entails a fundamental tension between statistical fidelity and computational tractability, as particle…
-
GenomeOcean Anywhere: Private WebGPU Inference for Genome MoEs
arXiv cs.LG
arXiv:2609.35882v2 Announce Type: replace-cross Abstract: Genome foundation models are most useful where sequences are generated, yet the largest models need datacenter accelerators and a place to send private DNA. We ask whether a…
-
TomoTransformer: Towards a Foundation Model for CT Reconstruction
arXiv cs.LG
arXiv:2609.37605v2 Announce Type: replace-cross Abstract: Supervised deep learning has advanced sparse-view tomographic reconstruction. However, conventional models, which typically map filtered back-projection (FBP) images or sinograms to…
-
THPL: A Vision-to-Language Decision Support Framework for Rainbow Trout Feeding Management in RAS
arXiv cs.AI
arXiv:2610.02378v1 Announce Type: new Abstract: In Recirculating Aquaculture Systems (RAS), precision feeding is critical for minimizing costs and improving fish welfare. However, existing methods lack cognitive alignment between fish…
-
MEA: A Reward-Driven Multi-Agent System for Faithful Model Explanations
arXiv cs.AI
arXiv:2610.02480v1 Announce Type: new Abstract: Recent years have seen the employment of a plethora of machine learning (ML) models in high-stakes domains, but they remain largely opaque to the practitioners who act on their predictions…
-
World Action Modeling with Progressive Visual Planning
arXiv cs.AI
arXiv:2610.02508v1 Announce Type: new Abstract: World action models (WAMs) have emerged as a promising paradigm for robotic control by jointly predicting future visual dynamics and actions from an initial observation and instruction…
-
Deep Bayesian REFoCUS
arXiv cs.LG
arXiv:2610.03419v1 Announce Type: new Abstract: In this work we formulate ultrasound multistatic recovery from arbitrary transmit sequences as a Bayesian inference problem. To that end, we train a deep generative prior on multistatic data…
-
Modeling Shared and Individual Structure for Cross-Subject Continuous Affect Regression from EEG-fNIRS
arXiv cs.AI
arXiv:2610.02796v1 Announce Type: new Abstract: Continuous, second-by-second valence-arousal estimation from physiological signals is typically studied in a subject-dependent setting, where the model sees labeled data from the same person…
-
Beyond Global Divergences: A Local-Mass Perspective on Bayesian Inference
arXiv cs.LG
arXiv:2606.27090v2 Announce Type: replace-cross Abstract: Global objectives, such as KL divergence and ELBO, are widely used in Bayesian inference for measuring distributional discrepancy. This paper studies distributional ``local-mass…
-
DriftWorld: Fast World Modeling through Drifting
arXiv cs.LG
arXiv:2607.15065v3 Announce Type: replace-cross Abstract: Predictive world models enable robots to simulate the visual outcomes of their actions, but state-of-the-art diffusion-based models remain costly because generating each rollout…
-
Learning to Price Electricity for Optimal Demand Response
arXiv cs.LG
arXiv:2610.00755v2 Announce Type: replace-cross Abstract: There is considerable interest in using time-varying electricity prices to shape consumer demand response, and better align energy demand with renewable production. However, optimal…
-
Large Language Continuous Diffusion Models
arXiv cs.CL
arXiv:2610.02665v1 Announce Type: new Abstract: Despite the success of discrete diffusion language models (dLMs) for fast parallel decoding, their non-smooth, high-dimensional space hinders trajectory steering for reasoning and inference…
-
Transcriptome-informed multi-modal AI for predicting neoadjuvant therapy response from breast cancer biopsies
arXiv cs.AI
arXiv:2610.03693v1 Announce Type: new Abstract: Scarcity of labeled data limits development of deep learning biomarkers in oncology. We develop a two-stage AI model predicting pathological complete response (pCR) to neoadjuvant therapy in…
-
ClarifyCodeBench: Evaluating LLMs on Clarifying Ambiguous Requirements for Code Generation
arXiv cs.AI
arXiv:2607.00711v2 Announce Type: cross Abstract: Large Language Models have emerged as programming assistants. However, the efficacy of code generation is constrained by the quality of input requirements, which are frequently ambiguous…
-
DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents
arXiv cs.AI
arXiv:2610.02320v1 Announce Type: cross Abstract: Computer-use agents need to reliably ground action targets in complex desktop scenes, where multiple applications, overlapping windows, and visually similar controls compete for attention…
-
MIRROR: Multipath Quorum Integrity for LLM Multi-Agent Communication
arXiv cs.AI
arXiv:2610.02349v1 Announce Type: cross Abstract: Inter-agent communication is central to Large Language Model Multi-Agent Systems (LLM-MAS), but it introduces an underexplored vulnerability: Agent-in-the-Middle (AiTM) attacks that…
-
Automating the Application of HCI Principles: Skills for On-Demand UI Construction, the Human-AI Space to Think, and the Future of HCI
arXiv cs.AI
arXiv:2610.02369v1 Announce Type: cross Abstract: Human-computer interaction (HCI) is in the middle of a transition: large language models can now generate functional user interfaces (UIs) on demand from natural-language task descriptions…
-
Student-Guided Teacher Distillation for Efficient LLM Task Routing: Positioning Against Jev-Style System-1 Classifiers
arXiv cs.LG
arXiv:2610.02516v1 Announce Type: new Abstract: Zero-shot classifiers are useful for routing user requests to specialized LLM tasks, but scoring every request against a large candidate set is expensive: a zero-shot NLI classifier must…
-
Prompted to Discriminate: Generalizing Malicious-Input Probes in the Wild
arXiv cs.LG
arXiv:2610.02413v1 Announce Type: new Abstract: LLM agents increasingly rely on activation probes as runtime monitors for prompt injection, jailbreaks, and unsafe requests, reading the model's own hidden state to catch a harmful input…
-
Keeping JEPA World Models Plannable When Little of the Frame Moves
arXiv cs.AI
arXiv:2610.03137v1 Announce Type: new Abstract: Specifying a goal in language rather than as a goal frame is a natural interface for planning with a latent world model, but testing it needs scenes in which language must discriminate…
-
Trained Agentic Context Management
arXiv cs.LG
arXiv:2610.02404v1 Announce Type: new Abstract: We study long context language models. Instead of training long context natively, or designing a long context harness, we train a model over the simplest possible harness: a tool to call…
-
Validated Data Onboarding for AI Demand Forecasting on U.S. Building Meter Data: Design, Controlled Evaluation, and a Corrected Negative Result
arXiv cs.LG
arXiv:2610.02397v1 Announce Type: new Abstract: Electric utilities and grid operators increasingly rely on machine-learning models to forecast next-day demand, and those models learn from meter data that is routinely defective: readings go…
-
CreateScore: Domain-Theory-Informed Bayesian Routing for LLM-Based CV Screening
arXiv cs.AI
arXiv:2610.02972v1 Announce Type: new Abstract: Large language models (LLMs) can support rubric-based screening of CVs, but applying a high-capability model to every candidate and criterion is costly. We present CreateScore, a…
-
Assessing Rule Adherence of LLM Adjudicators in Call of Cthulhu TRPG
arXiv cs.CL
arXiv:2607.02802v2 Announce Type: replace Abstract: As LLMs are increasingly deployed as autonomous adjudicators in games such as Call of Cthulhu (CoC), robust rule adherence becomes critical when user intent conflicts with system rules…
-
The Surprising Effectiveness of Shared Memory in Looped Transformers
arXiv cs.LG
arXiv:2610.02383v1 Announce Type: new Abstract: Looped Transformers apply the same layers several times per token, adding compute to improve quality without more parameters. Each recursion, however, writes its own key-value cache, so…
-
Parameter-Free Interval-Dynamic Regret under Heavy-Tailed Noise
arXiv cs.LG
arXiv:2610.02258v1 Announce Type: new Abstract: We study online convex optimization with one unbiased stochastic subgradient per round and an unknown finite conditional $p$th noise moment, $1
-
Positive-Unlabeled Learning for Agent Safety False Alarm Auditing
arXiv cs.AI
arXiv:2610.02925v1 Announce Type: new Abstract: Safety monitors help safeguard language-model agents interacting with external tools and environments, but conservative monitoring can generate many false alarms, consuming extensive review…
-
Overcoming Challenges of Interpretive Structural Modeling with Large Language Models
arXiv cs.LG
arXiv:2610.02254v1 Announce Type: new Abstract: Interpretive Structural Modeling (ISM) is a well-known process for multi-criteria decision making. The success of ISM over other methodologies is its ability to model causal relationships…
-
Morpheus: A Morphology-Aware Neural Tokenizer and Word Embedder for Turkish
arXiv cs.AI
arXiv:2606.18717v2 Announce Type: replace-cross Abstract: Turkish is agglutinative: meaning is carried by morphemes, yet the subword tokenizers that drive modern language models split words by corpus statistics, fragmenting semantically…
-
Nearest-neighbour baselines for fingerprint prediction from MS/MS spectra under different assumptions
arXiv cs.LG
arXiv:2610.02249v1 Announce Type: new Abstract: It has recently been shown that nearest-neighbour retrieval provides a strong baseline for molecular fingerprint prediction from MS/MS spectra, with several variants matching or outperforming…
-
FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training
arXiv cs.AI
arXiv:2610.02828v1 Announce Type: new Abstract: Adaptive LLM reinforcement-learning post-training changes multiple training actuators online, including rollout temperature, group size, clipping, KL regularization, verifier allocation, and…
-
Rhetorical Questions in LLM Representations: A Linear Probing Study
arXiv cs.CL
arXiv:2604.14128v3 Announce Type: replace Abstract: Rhetorical questions are asked not to seek information but to persuade or signal stance. How large language models internally represent them remains unclear. We analyze rhetorical…
-
Learning When to Commit from Partial Speech for End-to-End Simultaneous Speech Translation
arXiv cs.CL
arXiv:2610.02612v1 Announce Type: new Abstract: Simultaneous speech translation must emit useful target text before the source is complete while preserving every committed token. We adapt a full-utterance speech language model using prefix…
-
How Far Can You Get Without a GPU? A Systematic Benchmark of Lightweight Hallucination Detection Across Question Answering, Dialogue, and Summarisation
arXiv cs.AI
arXiv:2606.29809v2 Announce Type: replace-cross Abstract: Hallucination detection has become a pressing requirement for trustworthy AI deployment at scale. The most accurate detection methods depend on GPU-intensive inference, proprietary…
-
SovereignNegotiation-Bench: Evaluating User-Owned Personal Agents In Delegated Bargaining Under Privacy, Consent, Evidence, And Institutional Pressure
arXiv cs.AI
arXiv:2607.02814v2 Announce Type: replace-cross Abstract: Personal AI agents are beginning to negotiate for people, from refunds and bills to deposits and sales. A human agent in that position is judged by the duties owed to the principal…
-
Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts
arXiv cs.AI
arXiv:2607.19060v2 Announce Type: replace-cross Abstract: Fast prediction of the response of adhesive soft viscoelastic contacts represents a current challenge in soft robotics and for gripping and manipulation tasks. Determining the…
-
Embedded Bi-Temporal Building Damage Assessment for On-Board Data Reduction
arXiv cs.AI
arXiv:2609.37013v2 Announce Type: replace-cross Abstract: Rapid assessment of building damage after natural disasters is essential to support emergency response. Earth Observation satellites can acquire relevant imagery shortly after an…
-
MASCIT: A Mask-Aware State Space Classifier for Naturally Irregular Time Series
arXiv cs.AI
arXiv:2609.34409v2 Announce Type: replace-cross Abstract: Naturally irregular time series combine asynchronous observations, missing values, unequal lengths, and nonuniform sampling, while dense adapters can discard temporal structure. We…
-
Auditing Long-Term Memory Evaluation: Repeated Judging, Reader Variation, and Negative Controls
arXiv cs.AI
arXiv:2609.38021v2 Announce Type: replace-cross Abstract: This report audits evaluation of a long-term-memory retrieval chain on the 500 LongMemEval-S development questions. Its strongest historical reader lane scores 479 and 475 under an…
-
Safeguarding Mutual Correction in Source-Free Domain Adaptation via Cut Statistics
arXiv cs.AI
arXiv:2610.02981v1 Announce Type: new Abstract: Source-Free Domain Adaptation (SFDA) aims to adapt a source-pretrained model to an unlabeled target domain without access to the original source domain. While early single-model approaches…
-
Is a Picture Worth a Thousand Words? Adaptive Multimodal Fact-Checking with Visual Evidence Necessity
arXiv cs.AI
arXiv:2604.04692v3 Announce Type: replace-cross Abstract: Automated fact-checking is a crucial task that supports a responsible information ecosystem. While recent research has progressed from text-only to multimodal fact-checking, a…