Artificial Intelligence
Artificial Intelligence news — Monday 12 October 2026
150 stories from 9 sources, every one in English. Each links to its publisher.
-
Humanoid bottlenecks emerge amid US-China tensions
AI Business
Rising production volumes are increasing pressure on suppliers, while U.S.-China tensions could make it harder for Western manufacturers to secure alternatives to Chinese suppliers.
-
"The 3D machine that works with the word Han Mind"... "Anatomi for the clamping code" released
AI Times (Korea)
Translated from Korean
-
Ai2, MoE learning efficiency 2.7 times higher open source framework ‘Omo Core 3’ released
AI Times (Korea)
Translated from Korean
-
Sakana AI, LLM-based colleague review system with paper core claim error 73%
AI Times (Korea)
Translated from Korean
-
‘Data Center=Profit’...The reason for Antarpic in front of IPO
AI Times (Korea)
Translated from Korean
-
Can we change an AI model＆#8217;s architecture without retraining it from scratch?
Amazon Science Blog
AI agents are rapidly scaling: longer sessions, deeper reasoning chains, more autonomy. But the Transformer architecture underneath them still pays a memory tax that grows with the number of tokens it processes. As agents become more…
-
ORCARATOR, 100 million token context support Cyber Security Model Public
AI Times (Korea)
Translated from Korean
-
“AI Character Chatbot, Youth Protection Hole”. . Responsive Laws Responsive Tana
AI Times (Korea)
Translated from Korean
-
"AI instead of SI"...Silicon Valley Revenue on Trump Press
AI Times (Korea)
Translated from Korean
-
Game ·Delivery Companies Until...China AI model competition, re-explosion
AI Times (Korea)
Translated from Korean
-
[Bulletin board] KT 'Aodu AI' in front of the launch of the middle check, etc.
AI Times (Korea)
Translated from Korean
-
POSCO DX invests 20 billion in AI startups. ."Import AX Speed"
AI Times (Korea)
Translated from Korean
-
"Voice AI, not yet "ChatGPT Moment". . About Us
AI Times (Korea)
Translated from Korean
-
Why AI is Hard to Replace Human Counselors... “The limit to the end”
AI Times (Korea)
Translated from Korean
-
U.S. Government AI Regulated Blank Third-Party Evaluation Agency.. .'Establishment and Toxicity's Challenges
AI Times (Korea)
Translated from Korean
-
NAVER Maps, AI Eyeglasses Integration Guide Service Launch
AI Times (Korea)
Translated from Korean
-
“The clroad was conscious? ”...Andropic “AI Mysticism”
AI Times (Korea)
Translated from Korean
-
12~16 days Ginkgo Boiling Boiling Boiling Boiling Boiling Boiling Boiling
AI Times (Korea)
Translated from Korean
-
Yeongdeungpo-si, Gyeonggi-do, Republic of Korea
AI Times (Korea)
Translated from Korean
-
© Copyright 2017. All Rights Reserved.
AI Times (Korea)
Translated from Korean
-
Quality "AI Enterprises, 100B On-Device Model Drive Chip Requirements"
AI Times (Korea)
Translated from Korean
-
Prompt: AI agents are running into the limits of access
AI Business
As agents take on more work for users, companies are deciding whether to let outside AI systems act on their platforms.
-
Import AI 476: Creativity and RSI; worse-than-China open weight models; AI goes to space
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 New benchmark shows how AI models lack intuitive visual…
-
We Specified an AI Agent in 2016. We Called It Analytics.
Towards Data Science
Re-reading my own IBM patent on real-time concert feedback as what it actually was: an agentic loop with multimodal perception, memory-grounded inference, and a human in the loop. What it got right, and what a decade of hindsight teaches…
-
NAVER CLOUD, Vietnam FPT and Ata area AI factory expansion
AI Times (Korea)
Translated from Korean
-
We will introduce the "AI Leverage Evaluation System" to recruit employees and training.
AI Times (Korea)
Translated from Korean
-
USA, UK AI model pre-safe testing accepted.. .Trumf Ministry of Land, Infrastructure and Transport
AI Times (Korea)
Translated from Korean
-
The Download: moon city problems and drone traffic controllers
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. Mars may be a hellhole, but the moon is not far behind The US and China are racing to build bases on…
-
Amazon negotiating the acquisition of Decat 9 to strengthen AI infrastructure
AI Times (Korea)
Translated from Korean
-
EU and solar cybersecurity are discussed...Exhibition of high risk supplier restrictions
AI Times (Korea)
Translated from Korean
-
Han Electric KDN Introduces National Identity Verification Module for Solar Remote Control...
AI Times (Korea)
Translated from Korean
-
Australia, ‘Solar Housing’ Global Declaration Support... Double Up To 2030
AI Times (Korea)
Translated from Korean
-
I Built a Trap for AI Image Judges. They Walked Into a Different One.
Towards Data Science
Three vision AI models ignored the pretty image, yet GPT-4o invented answers and Claude kept answering B. The post I Built a Trap for AI Image Judges. They Walked Into a Different One. appeared first on Towards Data Science .
-
Mars may be a hellhole, but the moon is not far behind
MIT Technology Review Biotech
Welcome to the second space race. You can be forgiven for experiencing déjà vu. Last time, it was the Soviet Union and the United States vying to be the first to plant some flags. Now it’s a sprint to the lunar surface between America and…
-
[AIReader's Book] Hands-on Machine Learning with Simone. Skills
AI Times (Korea)
Translated from Korean
-
The ecologist revealing the invisible networks that revive rainforests
MIT Technology Review Biotech
In a research station in central Panama, you can walk through a century of forest time, watching scrubby plots of adolescent trees give way to cathedral-like canopies of mature growth. Forest ecologist Daisy Dent, a senior scientist at…
-
Golden Anchor for Video Generation
arXiv cs.LG
arXiv:2610.13217v1 Announce Type: cross Abstract: Video generation supports digital content creation, with applications in virtual avatars, AI-driven filmmaking, interactive game engines, and world models. State-of-the-art systems achieve…
-
Not All Dimensions Are Equal: Outlier-Aware Attention Linearization with Grouped Quadratic Kernels
arXiv cs.AI
arXiv:2610.13674v1 Announce Type: cross Abstract: Linearizing softmax attention offers a route to efficient Transformer inference, but preserving pretrained attention remains challenging. We identify an important source of approximation…
-
Workhorse: Learning Robust Whole-Body Humanoid Loco-Manipulation from Human Data
arXiv cs.LG
arXiv:2610.09117v2 Announce Type: replace-cross Abstract: Humanoid robots still struggle to plan contact-rich whole-body manipulation from egocentric RGB and proprioception. Workhorse learns such manipulation from robot-free human…
-
AI Models Can Provably Hide Arbitrary Capabilities
arXiv cs.LG
arXiv:2610.13928v1 Announce Type: new Abstract: Capability evaluations aim to surface dangerous model behaviors before deployment, but their reliability depends on the assumption that hidden capabilities can be elicited. We show this…
-
Fear Is Not Always Fear: Variation in Emotion Label Equivalence Across Datasets
arXiv cs.CL
arXiv:2610.12841v1 Announce Type: new Abstract: Emotion classifiers are routinely applied beyond their training data, on the assumption that same-named labels such as fear or surprise mean the same thing across resources. We test this…
-
AMLworld-Compact: Importance-Weighted Downsampling for Cost-Effective LLM Evaluation and Error Diagnosis
arXiv cs.CL
arXiv:2610.13627v1 Announce Type: new Abstract: Evaluating large language models (LLMs) on AMLworld's smallest test splits is costly because each of over a million targets requires a separate graph prompt. Sampling reduces this cost, but…
-
A Unified Benchmark for Dynamic Medical Treatment Reinforcement Learning
arXiv cs.LG
arXiv:2606.01028v4 Announce Type: replace Abstract: Medical treatment recommendation poses several challenges to reinforcement learning (RL): patient physiology evolves in continuous time, measurements and interventions are performed at…
-
DualQuant: Dual-scale Calibration-free Weight-and-Activation Quantization for LLMs
arXiv cs.AI
arXiv:2610.13767v1 Announce Type: cross Abstract: Low-bit inference on accelerators with native 4-bit support requires quantizing both weights and activations to realize the full throughput gains these formats provide. While post-training…
-
Inject a Turn, Not a Token: Learning Where to Redirect LLM Reasoning for Self-Correction
arXiv cs.AI
arXiv:2610.13672v1 Announce Type: cross Abstract: Large language models struggle to reliably identify and correct errors in their own reasoning. Unlike prior training-based approaches that fine-tune an LLM to generate revisions or…
-
Representation Alignment from a Few Paired Data via Distilled Optimal Transport
arXiv cs.AI
arXiv:2610.12781v1 Announce Type: new Abstract: Aligning two independently pretrained unimodal encoders typically demands millions of paired examples. In this paper, we study the regime where only $N \sim 100$--$500$ pairs are available…
-
Looking Under the Streetlight: Evaluation in Generative Molecular Dynamics
arXiv cs.LG
arXiv:2610.12518v1 Announce Type: cross Abstract: This position paper argues that the generative molecular dynamics (GenMD) community is not properly evaluating its progress: current evaluations let a model match what is reported while…
-
X-AgentRed: Red-Teaming Personal Agents in Evolving Multi-Service Ecosystems
arXiv cs.AI
arXiv:2610.13886v1 Announce Type: new Abstract: Personal agents increasingly connect to services across the web and applications by using tools. As the effects of their actions accumulate and interact across services, unintended or harmful…
-
Hide and Seek: HIDE Concepts to SEEK explanations for black box models
arXiv cs.LG
arXiv:2610.13935v1 Announce Type: cross Abstract: Interpreting machine learning model predictions is a central challenge, particularly for black-box vision models. Concept-based attribution methods interpret models using…
-
LongHarness: Advancing Agentic Reasoning over 100M-Token Corpora via Stateful Evidence Management
arXiv cs.AI
arXiv:2610.13075v1 Announce Type: new Abstract: Corpus-level long-context reasoning requires agents to acquire and integrate evidence across document collections that exceed their context windows. Existing agent harnesses enable…
-
Bridging Human and Model Attention: Explainability Analysis of CNN, Mamba, and ViT Architectures with Gaze-Based Validation
arXiv cs.LG
arXiv:2610.12703v1 Announce Type: cross Abstract: High-risk applications of deep learning require precise characterization of model decision-making. We conduct a systematic explainability evaluation across Convolutional Neural Networks…
-
Low-Rank Kolmogorov-Arnold Generator Flows for Interpretable Operator Learning
arXiv cs.LG
arXiv:2610.13028v1 Announce Type: new Abstract: Neural operators that learn physical laws from data should generalize beyond their training distribution and expose the learned law in closed form. The Kolmogorov-Arnold neural operator…
-
How Much Computation Does Every Forecast Need? Structured Block Skipping in Two Time-Series Foundation Models
arXiv cs.LG
arXiv:2610.13208v1 Announce Type: new Abstract: Time-series foundation models use the same network depth for every input, although forecasting windows differ in structure and difficulty. We test whether whole Transformer blocks can be…
-
Trainable Quantum-Native Language Models: Dissipative Memory, Capacity Limits, and Hardware Validation
arXiv cs.LG
arXiv:2610.12806v1 Announce Type: cross Abstract: We study byte-level autoregressive language models whose hidden state is a small quantum register, whose transitions are learned token-conditioned quantum channels, and whose next-byte…
-
From Imbalance to Evidence: Deep Nonparametric Testing for Complex Data under Severe Sample Imbalance
arXiv cs.LG
arXiv:2610.12946v1 Announce Type: cross Abstract: Comparing distributions under severe sample imbalance requires effectively using abundant majority observations while retaining sensitivity to differences involving the smaller sample. This…
-
JoyInside-VLM-Agent: Product-Conditioned Multimodal Policy Learning for Embodied Agents
arXiv cs.LG
arXiv:2610.12957v1 Announce Type: cross Abstract: Embodied agents must translate multimodal observations and user requests into executable actions while adapting to product constraints, changing goals, and execution feedback. We present…
-
Scan-Parallel Variational Inference for event-driven Latent Jump Models
arXiv cs.LG
arXiv:2610.12986v1 Announce Type: cross Abstract: Continuous-time latent state-space models are widely used for irregularly observed time series, typically representing uncertainty through a random initial state followed by smooth ordinary…
-
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training
arXiv cs.LG
arXiv:2610.12995v1 Announce Type: cross Abstract: Supervised fine-tuning (SFT) remains a widely used approach for post-training large language models, because fixed demonstrations can directly introduce desired reasoning strategies and…
-
Lightweight Time-Frequency Vibration Sensing for Consumer Device Condition Monitoring
arXiv cs.LG
arXiv:2610.13040v1 Announce Type: cross Abstract: Consumer rotating devices, including washing machines, air conditioners, electric fans, robotic vacuum cleaners, and home fitness equipment, increasingly rely on vibration sensing for…
-
CoBe: A Benchmark For Conversational Counterfactual Text Editing
arXiv cs.LG
arXiv:2610.13057v1 Announce Type: cross Abstract: Counterfactual text editing is the ability to revise a given text in light of a hypothetical change while preserving facts not causally affected by that change. When asked what would have…
-
OGPO: Offline Goal-conditioned Policy Optimization for Recoverable Vision-Language-Action Models
arXiv cs.AI
arXiv:2610.12914v1 Announce Type: new Abstract: Vision-language-action (VLA) models have shown strong promise in robotic manipulation, yet their post-training still predominantly relies on trajectory-level supervised imitation over robot…
-
Masked Diffusion Language Agents for Tool-Integrated Chemical Reasoning
arXiv cs.LG
arXiv:2610.13330v1 Announce Type: cross Abstract: Designing tool-integrated reasoning agent systems for complex chemical tasks remains a fundamental challenge. Although recent chemical agents have demonstrated promising performance through…
-
RLVR$^{2}$: Reinforcement Learning with Verifiable Rubric-based Ranking
arXiv cs.AI
arXiv:2609.23457v2 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) is expanding from tasks with well-defined correctness signals, such as mathematics and code, toward multifaceted quality…
-
A Multi-Agent LLM System for Automated Tuning of Specialized Neural Network Models
arXiv cs.LG
arXiv:2610.13663v1 Announce Type: cross Abstract: Many machine learning approaches are two-level: an expert selects a class of methods, within this class the choice of key components is critical and usually manual. This paper addresses the…
-
Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study
arXiv cs.LG
arXiv:2610.13287v1 Announce Type: cross Abstract: Prompt optimizers rewrite a prompt to raise one target metric, but prompts also state requirements the metric does not score, such as an output format, citations, or abstention. In a…
-
A Distributional Framework for Behavioral Characterization and Risk Management of Large Language Models
arXiv cs.LG
arXiv:2610.12718v1 Announce Type: cross Abstract: Large language models are stochastic by nature. Repeated executions under the same experimental conditions may produce different responses. We propose a framework that can be applied across…
-
Specialists Learn Faster: Knowledge Organization in Pre-training Shapes Knowledge Update Efficiency
arXiv cs.LG
arXiv:2610.13835v1 Announce Type: new Abstract: An ideal (language) model must continually update its knowledge. Existing work largely focuses on making these updates efficient after training. We investigate how the organization of…
-
IAD-Unify: Task-Specific Interfaces for Industrial Anomaly Understanding, Segmentation, and Generation
arXiv cs.AI
arXiv:2604.12440v3 Announce Type: replace-cross Abstract: Industrial anomaly inspection requires complementary capabilities: explaining an observed defect, localizing its pixels, and synthesizing a controlled edit. We present IAD-Unify, a…
-
A Theory of the Mechanics of Information: Generalization Through Measurement of Uncertainty (Learning is Measuring)
arXiv cs.AI
arXiv:2510.22809v2 Announce Type: replace-cross Abstract: Traditional machine learning relies on explicit models and domain assumptions, often limiting flexibility and interpretability. We introduce a "model-free" framework of…
-
RAMP: Recurrent Adaptive Mixed Precision for Efficient Prefill in Recurrent-Depth Language Models
arXiv cs.CL
arXiv:2610.13041v1 Announce Type: new Abstract: Recurrent-depth language models improve inference-time reasoning by repeatedly applying a shared network to refine internal representations, but the additional recurrent computation also…
-
Auditing Prompt Injection Defenses Before They Reach Users: What Strong Adversaries Reveal
arXiv cs.AI
arXiv:2505.15738v3 Announce Type: replace-cross Abstract: Large language models are increasingly deployed in user-facing products that handle sensitive personal data while interacting with untrusted external content. This makes them a…
-
Your Discrete Graph Diffusion Model is Secretly Memorizing
arXiv cs.AI
arXiv:2610.13786v1 Announce Type: cross Abstract: Discrete graph diffusion models (GDMs) corrupt the graph structure by adding noise independently to nodes and edges under a Markovian noise model. In the unconditional setting without node…
-
RLVR$^{2}$: Reinforcement Learning with Verifiable Rubric-based Ranking
arXiv cs.LG
arXiv:2609.23457v2 Announce Type: replace Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) is expanding from tasks with well-defined correctness signals, such as mathematics and code, toward multifaceted quality requirements…
-
Variance-Reduced Policy Gradient is Nearly Minimax-Optimal
arXiv cs.LG
arXiv:2610.13692v1 Announce Type: new Abstract: Standard Policy Gradient (PG) algorithms suffer from high estimation variance because they rely on Monte Carlo updates. While off-policy variance reduction mitigates this problem…
-
The Evidence Anchoring Boundary: Where Visual Interventions Matter and When They Help in Multimodal LLMs
arXiv cs.AI
arXiv:2610.12981v1 Announce Type: cross Abstract: Visual interventions provide a training-free way to reduce hallucinations in multimodal large language models. However, an intervention may have little leverage at some layers and may…
-
Medical foundation models converge less under label supervision
arXiv cs.LG
arXiv:2607.20274v3 Announce Type: replace-cross Abstract: Diagnostic classifiers and imaging biomarkers are fitted on the embeddings of medical foundation models. These models are replaced as new versions appear. This practice assumes that…
-
BiGym 2.0: Benchmarking Learned and Agent-Developed Policies for Humanoid Household Manipulation
arXiv cs.LG
arXiv:2610.07594v2 Announce Type: replace-cross Abstract: Humanoid household manipulation requires the arms to act while the body balances, steps and changes posture. We present BiGym 2.0, an adaptation of BiGym for the Unitree G1 across…
-
Cross-View Attention Fusion Net: A Prior-Guided Dual-View Representation Learning for Cardiac Output Estimation from Short-Term PPG Signals
arXiv cs.LG
arXiv:2605.19666v2 Announce Type: replace-cross Abstract: Accurate cardiac output (CO) estimation from photoplethysmography (PPG) is promising for unobtrusive hemodynamic monitoring, but remains difficult since CO is jointly determined by…
-
UxSID: Semantic-Aware User Interests Modeling for Ultra-Long Sequence
arXiv cs.LG
arXiv:2605.09040v4 Announce Type: replace-cross Abstract: Modeling ultra-long user sequences involves a difficult trade-off between efficiency and effectiveness. While current paradigms rely on either item-specific search or item-agnostic…
-
Expert-level detection of low-voltage intervals in EEG as a robust biomarker of neonatal encephalopathy
arXiv cs.LG
arXiv:2610.12680v1 Announce Type: new Abstract: Background: Discontinuity in neonatal EEG is a critical marker of brain dysfunction but can be difficult to quantify visually. Inter-burst intervals (IBIs) are a key characteristic of…
-
FactoryBench: Evaluating Industrial Machine Understanding
arXiv cs.LG
arXiv:2605.07675v2 Announce Type: replace-cross Abstract: We introduce FactoryBench, a benchmark for evaluating time-series models and LLMs on machine understanding over industrial robotic telemetry. Q&A pairs are organized along four…
-
Component-Adaptive and Lesion-Level Supervision for Improved Small Structure Segmentation in Brain MRI
arXiv cs.LG
arXiv:2604.08015v4 Announce Type: replace-cross Abstract: Small lesions in brain MRI are hard to segment because they occupy a tiny fraction of the volume and are dominated by background and larger lesions during voxel-wise optimization…
-
Bistable by Construction: Wall-Clock-Calibrated State Monitors Have No Moment-Detection Regime at Agent Cadence
arXiv cs.LG
arXiv:2606.19386v2 Announce Type: replace-cross Abstract: Runtime monitors for autonomous agents commonly threshold an accumulated internal state - a behavioural baseline, a drift statistic, or, in our prior work, a modelled affective…
-
DRIFT: Derailing Denoising Trajectories of Flow-Matching VLAs with Adversarial Patch Attack
arXiv cs.LG
arXiv:2608.03207v2 Announce Type: replace-cross Abstract: Flow-matching vision-language-action (VLA) models such as pi0 generate robot actions by integrating a learned denoising velocity field, and have been reported to resist adversarial…
-
Enhancing Tabular Learners with Context-Aware Semantic Embeddings
arXiv cs.LG
arXiv:2608.03565v2 Announce Type: replace-cross Abstract: While modern tabular learners excel at capturing statistical patterns, they frequently operate in a semantic vacuum, treating textual features as discrete symbols, ignoring the rich…
-
Nonsmooth Optimization via Orthogonalized Momentum
arXiv cs.LG
arXiv:2609.13677v2 Announce Type: replace-cross Abstract: Modern real application problems involve matrix-valued parameters, yet conventional optimizers treat them as vectors, thereby motivating matrix-aware methods that exploit…
-
An Input-Dependent Readout Phase for Resonant Sequence Models
arXiv cs.LG
arXiv:2610.12496v1 Announce Type: new Abstract: Sequence models with a fixed-size state avoid the quadratic cost of attention, but face a design choice. An input-dependent state transition gives content-dependent memory at the cost of…
-
Coordinate-Free Automatic Differentiation
arXiv cs.LG
arXiv:2610.12776v1 Announce Type: new Abstract: Automatic differentiation (AD) is a cornerstone of modern machine learning. Since all AD frameworks implicitly fix a basis of Euclidean space to define gradients and Jacobians, they natively…
-
MolP2SBench: Do LLMs Really Reason from Molecular Properties to Structure?
arXiv cs.LG
arXiv:2610.12522v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used in molecular discovery and structure elucidation, where decisions depend on physical observations. Existing evaluations primarily focus on…
-
SqueezeBench: Saturation-Resilient Evaluation of Mathematical Capability
arXiv cs.LG
arXiv:2610.13461v1 Announce Type: cross Abstract: AI systems are making rapid progress in mathematics, but this progress is hard to measure. Benchmarks of solved problems saturate, and in benchmarks of open problems each problem can be…
-
Vocal Intensity Conversion with Continuous SPL Control via Adversarial Training
arXiv cs.LG
arXiv:2610.13608v1 Announce Type: cross Abstract: Vocal intensity conversion renders a recorded utterance at a different vocal intensity, rather than simply scaling its amplitude, while preserving its content and speaker identity. Existing…
-
Electron-Density-Informed Molecular Property Prediction Using Surrogate Functionals
arXiv cs.LG
arXiv:2610.13649v1 Announce Type: cross Abstract: Machine learning methods are increasingly used to predict chemical properties directly from the molecular structure to bypass the often prohibitive cost of quantum mechanical calculations…
-
VLA-FAIL: Efficient Task Failure Detection for Finetuned Vision-Language-Action Models
arXiv cs.LG
arXiv:2606.21386v2 Announce Type: replace Abstract: Vision-language-action models (VLAs) achieve state-of-the-art performance on many robotic manipulation tasks, yet they can still behave unpredictably in out-of-distribution scenarios…
-
Approximation of differential entropy in Bayesian optimal experimental design
arXiv cs.LG
arXiv:2510.00734v2 Announce Type: replace-cross Abstract: Bayesian optimal experimental design provides a principled framework for selecting experimental configurations that maximize obtained information. In this work, we consider settings…
-
ABLEIST: Intersectional Disability Bias in LLM-Generated Hiring Scenarios
arXiv cs.LG
arXiv:2510.10998v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly under scrutiny for perpetuating identity-based discrimination in high-stakes domains such as hiring, particularly against people with…
-
RanSOM: Second-Order Momentum with Randomized Scaling for Constrained and Unconstrained Optimization
arXiv cs.LG
arXiv:2602.06824v3 Announce Type: replace-cross Abstract: Second-order momentum corrects the lag of momentum behind the current gradient with a Hessian-vector product (HVP), but existing corrections need either a Lipschitz Hessian or extra…
-
Lesion-DDPM: Lesion-Enhanced 3D Diffusion for MS MRI Synthesis
arXiv cs.LG
arXiv:2606.15457v2 Announce Type: replace-cross Abstract: 3D FLAIR MRI is widely recommended as one of the standard MRI sequences for brain imaging in multiple sclerosis (MS), but publicly available MS datasets remain relatively small and…
-
Democratizing MoE inference on commodity GPUs with CoMoE
arXiv cs.LG
arXiv:2610.09424v2 Announce Type: replace-cross Abstract: Deploying Mixture-of-Experts (MoE) models relies heavily on Expert Parallelism, which generates intense inter-GPU communication. Consequently, state-of-the-art inference systems…
-
Explicit Asymptotic Bounds for Sequential Calibration Beyond $T^{2/3}$
arXiv cs.LG
arXiv:2610.07623v2 Announce Type: replace-cross Abstract: Probability forecasts are calibrated when predicted probabilities match empirical outcome frequencies: among events assigned a probability $p$, we'd hope that the fraction of…
-
Machine Learning and Time-Series Foundation Models for Day-Ahead Electricity Forecasting in a University Building
arXiv cs.AI
arXiv:2610.12638v1 Announce Type: cross Abstract: Day-ahead electricity forecasts can help building operators schedule heating, ventilation, and other services around anticipated demand. Choosing a forecasting model also requires balancing…
-
Trajectory Prediction for Autonomous Driving in Road Construction Zones Using Human-in-the-Loop Driving Data
arXiv cs.AI
arXiv:2610.12747v1 Announce Type: cross Abstract: Ego-trajectory prediction in urban roadway work zones remains an underexplored topic in autonomous-driving research, with limited dedicated data and systematic evaluation. Roadway work…
-
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding
arXiv cs.AI
arXiv:2610.12939v1 Announce Type: cross Abstract: Decoder hidden states from different layers of large language models (LLMs) often encode complementary information. However, existing speculative decoding methods typically rely on a small…
-
Pretraining and Test-Time Training Budgets Matter for ARC: A Controlled Study of VARC and TRM
arXiv cs.AI
arXiv:2610.13660v1 Announce Type: cross Abstract: The Abstraction and Reasoning Corpus (ARC) is increasingly used to evaluate whether neural networks can infer task-specific transformations from few demonstrations. Recent ARC models differ…
-
Estimating Great Expectations Under Autoregressive Language Models with Potentials
arXiv cs.AI
arXiv:2610.11399v2 Announce Type: replace Abstract: Many applications of language models hinge not on individual samples but on the expectation of a test functional under the model. Estimating such expectations reliably can be…
-
PACT: Privileged Trace Co-Training for Multi-Turn Tool-Use Agents
arXiv cs.AI
arXiv:2606.16215v2 Announce Type: replace-cross Abstract: Multi-turn tool-use agents must reason, call tools, and adapt to observations across several interaction turns. Post-training such agents is challenging, as reinforcement learning…
-
Video2World: Benchmarking Coding Agents for Interactive World Modeling from Embodied Videos
arXiv cs.AI
arXiv:2610.04432v3 Announce Type: replace-cross Abstract: Building interactive simulators from real-world observations is a promising way to scale embodied data, but current pipelines still rely heavily on manual environment construction…
-
HART: Hybrid Attention-RWKV Time series Foundation Model
arXiv cs.LG
arXiv:2610.12883v1 Announce Type: new Abstract: Non-autoregressive multi-step forecasting (NAF) enables time series foundation models to predict an entire future horizon in a single forward pass. Despite its strong sequence modeling…
-
Heterogeneity-Aware Client Sampling for Optimal and Efficient Federated Learning
arXiv cs.LG
arXiv:2505.11304v4 Announce Type: replace Abstract: Federated learning (FL) commonly involves clients with diverse communication and computational capabilities. Such heterogeneity can significantly distort the optimization dynamics and…
-
asdex: Automatic Sparse Differentiation in JAX
arXiv cs.LG
arXiv:2610.12336v2 Announce Type: replace-cross Abstract: Many tasks in scientific computing and machine learning require the Jacobian or Hessian matrix of a function. Automatic differentiation (AD) computes these derivatives to machine…
-
The Predictability Bias of Joint-Embedding Predictive Learning in Medical Imaging
arXiv cs.LG
arXiv:2610.13237v1 Announce Type: cross Abstract: Joint-embedding predictive architectures (JEPAs) learn by predicting teacher representations of masked regions from visible context. Point prediction summarizes the conditional target…
-
Mixture of Many Zero-Compute Experts: A High-Rate Quantization Theory Perspective
arXiv cs.LG
arXiv:2510.03151v2 Announce Type: replace Abstract: This paper uses classical high-rate quantization theory to provide new insights into mixture-of-experts (MoE) models for regression tasks. Our MoE is defined by a segmentation of the…
-
A Theory of the Mechanics of Information: Generalization Through Measurement of Uncertainty (Learning is Measuring)
arXiv cs.LG
arXiv:2510.22809v2 Announce Type: replace Abstract: Traditional machine learning relies on explicit models and domain assumptions, often limiting flexibility and interpretability. We introduce a "model-free" framework of instance-based…
-
When Should the Teacher Move? Temporal Coupling and Stability in Self On-Policy Distillation
arXiv cs.LG
arXiv:2606.03532v2 Announce Type: replace Abstract: Self on-policy distillation trains a student policy against a teacher derived from its own parameter history, yet the teacher's update schedule -- which governs the \emph{temporal…
-
Rehearse Everything, Remember Nothing: Attic-KV Rehearses What Will Be Read
arXiv cs.LG
arXiv:2610.12133v2 Announce Type: replace-cross Abstract: Many key-value (KV) caches are compressed before anyone knows what will be asked of them: a document cached for retrieval, a prompt prefix shared across requests, the memory of a…
-
Equal-Budget Monte Carlo Tree Search versus Randomized-Restart List Scheduling for DAG Workflows: A Controlled Equal-Budget Computational Study
arXiv cs.AI
arXiv:2610.12542v1 Announce Type: new Abstract: Monte Carlo Tree Search (MCTS) is often proposed for scheduling directed-acyclic-graph (DAG) workflows, yet most studies never run the comparison that could refute it - giving a much simpler…
-
KITE: A Knowledge-Injected Temporal Encoder for Time-Series Event Forecasting
arXiv cs.AI
arXiv:2610.12738v1 Announce Type: cross Abstract: Forecasting rare, high-cost events in multivariate time-series systems is hard because their precursors are seldom observed, so a purely data-driven model has little to learn from. The…
-
Continually Post-training Small Models via Synthetic Self-distillation
arXiv cs.AI
arXiv:2610.13113v1 Announce Type: cross Abstract: Continual post-training faces a dual challenge of acquisition and retention: models must learn new capabilities while preserving prior knowledge. While Self-Distillation Fine-Tuning (SDFT)…
-
Making LLMs Say What They Think: Measuring and Improving CoT-Interpretability Alignment
arXiv cs.AI
arXiv:2609.38972v2 Announce Type: replace-cross Abstract: Chain-of-thought (CoT) traces often serve as a proxy for how Large Language Models (LLMs) arrive at their answers. However, growing evidence shows that models' CoT often fails to…
-
EvoBelief: Falsifiable World Models for Self-Improving LLM Agents
arXiv cs.AI
arXiv:2610.12789v1 Announce Type: new Abstract: Self-evolving LLM agents typically improve how they act through accumulated experience, reflection, or skills. Yet even procedurally capable agents can fail when their beliefs about…
-
DualQuant: Dual-scale Calibration-free Weight-and-Activation Quantization for LLMs
arXiv cs.LG
arXiv:2610.13767v1 Announce Type: new Abstract: Low-bit inference on accelerators with native 4-bit support requires quantizing both weights and activations to realize the full throughput gains these formats provide. While post-training…
-
Neural networks are more modular than single neurons suggest
arXiv cs.AI
arXiv:2610.12842v1 Announce Type: new Abstract: Neural networks routinely develop representations in which individual neurons contribute to multiple distinct functions, leading to the widespread view that such networks lack modular…
-
Mind the Spatial Gap: A Benchmark for Geospatial Reasoning and Constrained Spatial Editing
arXiv cs.AI
arXiv:2610.12881v1 Announce Type: new Abstract: Understanding and manipulating spatial relations is fundamental to AI systems that reason about physical environments and construct grounded world models. However, existing evaluations of…
-
Adaptive Variance Reduction without Knowing Oracle Smoothness
arXiv cs.LG
arXiv:2610.13799v1 Announce Type: cross Abstract: Variance reduction improves the convergence rate of stochastic non-convex optimization under the average smoothness condition. Existing adaptive variance reduction methods are especially…
-
Teaching neural decoders with language
arXiv cs.AI
arXiv:2610.12921v1 Announce Type: new Abstract: Brain-computer interfaces decode neural activity, yet class-label supervision (CLS) leaves relationships among classes implicit. We propose Language as Supervision (LaS), which aligns neural…
-
MathVis: Can AI Reason through Drawing? A Benchmark for Mathematical Figure Generation
arXiv cs.AI
arXiv:2610.12954v1 Announce Type: new Abstract: This paper introduces MathVis, a benchmark designed to evaluate LLMs' ability to visualize frontier mathematical problems and perform diagram-assisted reasoning. The benchmark consists of 512…
-
AI-Mediation in Retail Markets: Revenue Effects and Assortment Optimization
arXiv cs.AI
arXiv:2610.12917v1 Announce Type: new Abstract: AI shopping assistants increasingly influence which products consumers consider. We study retail assortment planning under AI mediation, where a retailer chooses a catalog, an AI intermediary…
-
Convex Optimization of Spiking Neural Networks with Global and Approximation Guarantees
arXiv cs.LG
arXiv:2605.08022v3 Announce Type: replace-cross Abstract: Spiking neural networks (SNNs) have been proposed as energy-efficient alternatives to deep neural networks. Their training is non-convex and involves a non-differentiable spike…
-
Amortised Inference through One-Step Implicit Sampling
arXiv cs.LG
arXiv:2610.12679v1 Announce Type: new Abstract: Efficient sampling from Bayesian posterior distributions in high dimensions remains a hard problem, despite advances in theory and algorithms. Multiple steps are usually required to draw…
-
A Generalizable Machine Learning Framework for Half-Heusler Thermoelectrics
arXiv cs.LG
arXiv:2602.01149v2 Announce Type: replace-cross Abstract: Machine learning (ML) has emerged as a powerful approach for accelerating the discovery of thermoelectric (TE) materials, yet many models suffer from limited generalizability. Here…
-
UniConFlow: A Unified Constrained Flow-Matching Framework for Certified Motion Planning
arXiv cs.LG
arXiv:2506.02955v3 Announce Type: replace-cross Abstract: Generative models have become increasingly powerful tools for robot motion generation, enabling flexible and multimodal trajectory generation across various tasks. Yet, most…
-
StereoEngine: Scaling Stereo Bimanual Robot Data without Stereo Cameras
arXiv cs.LG
arXiv:2610.12734v1 Announce Type: cross Abstract: Stereo vision is increasingly used for robotic manipulation, with a growing body of work incorporating it into Vision-Language-Action (VLA) models. However, the scarcity of large-scale…
-
Fast localized urban microclimate emulator: wind and temperatures in unseen urban morphologies
arXiv cs.LG
arXiv:2503.19708v3 Announce Type: replace-cross Abstract: Urban microclimate, encompassing wind and temperature fields shaped by building geometry, strongly affects building energy use, pedestrian wind comfort, pollutant dispersion, the…
-
When Do External Priors Help Low-Resource Structured Prediction? A Ranking-Level Alignment Study
arXiv cs.LG
arXiv:2610.11146v2 Announce Type: replace Abstract: Cold-start credit scoring (scarce labels, immature features, or minimal capacity) motivates injecting external priors (teacher models, LLMs, or domain experts) into training. We study…
-
Toward Maximizing Algebraic Connectivity via Density-Dependent Backbone Construction and Reinforcement Learning Completion
arXiv cs.LG
arXiv:2610.12697v1 Announce Type: cross Abstract: This paper studies an open network design problem of constructing graph topologies with maximum algebraic connectivity under fixed numbers of vertices and edges. A density-dependent…
-
Instance-Level Difficulty: A Missing Perspective in Machine Unlearning
arXiv cs.LG
arXiv:2410.03043v3 Announce Type: replace Abstract: Current research on deep machine unlearning primarily focuses on improving or evaluating the overall effectiveness of unlearning methods while overlooking the varying difficulty of…
-
Work While They Sleep: Exploiting Evaluation Latency for Fully Bayesian Optimization
arXiv cs.LG
arXiv:2610.08969v2 Announce Type: replace Abstract: Black-box optimization problems are ubiquitous across science and engineering, often dealing with expensive objective functions. This objective latency has two consequences during…
-
LLM4Trust: Exploring the Capabilities of Large Language Models for Trust Evaluation
arXiv cs.LG
arXiv:2609.33521v2 Announce Type: replace Abstract: Trust evaluation plays a critical role in cybersecurity by supporting risk mitigation and decision-making. A variety of trust evaluation methods have been proposed, with learning-based…
-
AutoIndexer: Flexible-Order Decoding by Training Causal Models on Chains of Edits
arXiv cs.LG
arXiv:2610.12545v1 Announce Type: cross Abstract: Autoregressive decoding is less suited to non-causally ordered data, e.g., cross-references or future dependencies, since their outputs cannot be revised post hoc, unlike iterative human…
-
Protocol effects on feature-based hardware-Trojan detection across Trust-Hub families
arXiv cs.LG
arXiv:2609.07199v2 Announce Type: replace Abstract: Trust-Hub reuses host circuits: several files differ mainly in the inserted Trojan. When gates from sibling variants enter both training and test folds, a detector can benefit from host…
-
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training
arXiv cs.AI
arXiv:2610.12995v1 Announce Type: cross Abstract: Supervised fine-tuning (SFT) remains a widely used approach for post-training large language models, because fixed demonstrations can directly introduce desired reasoning strategies and…
-
Co-Modifying Application Code and Deployment Infrastructure: A Causal Study of AI Agent Pull Requests
arXiv cs.AI
arXiv:2610.12849v1 Announce Type: cross Abstract: Current software engineering agents are evaluated predominantly inside isolated sandboxes against synchronous local tests. However, deploying features in collaborative repositories requires…
-
DASC: Decay-Aware State Compression for Hybrid Linear-Attention Serving
arXiv cs.LG
arXiv:2608.30386v2 Announce Type: replace Abstract: Hybrid linear-attention architectures have recently scaled to large open-weight models, offering quality competitive with full attention while substantially reducing key/value (KV) cache…
-
Private and Distributionally Robust Federated Multiclass Classification
arXiv cs.LG
arXiv:2610.13631v1 Announce Type: new Abstract: Fault diagnosis in asset-intensive industries requires multiclass models trained on noisy, sensitive data collected across geographically distributed client sites. Federated learning enables…
-
Does Location Matter? A Black-Box Causal Audit Reveals Model-Dependent Audio Grounding in LALM Hallucination
arXiv cs.LG
arXiv:2610.13112v1 Announce Type: cross Abstract: Large audio-language models (LALMs) frequently exhibit object hallucination, affirming absent sounds. The dominant mitigation, Audio-Aware Decoding (AAD), contrasts logits under real audio…
-
MagicBeanStalk: Escaping the RLVR Zero Advantage Trap with Recursive Data Synthesis
arXiv cs.LG
arXiv:2610.12659v1 Announce Type: cross Abstract: Data-centric scaling of RLVR has focused on collecting or generating verifiable training problems. However, corpus size alone does not determine whether the dataset provides a useful…
-
Incremental Retrieval-Augmented Generation for Human-in-the-Loop Decision Support
arXiv cs.AI
arXiv:2610.12666v1 Announce Type: cross Abstract: We introduce iRAG, a human-in-the-loop decision-support system combining Retrieval-Augmented Generation (RAG) with Hybrid Decision-Making. Unlike conventional RAG systems, relying on static…
-
CoBe: A Benchmark For Conversational Counterfactual Text Editing
arXiv cs.AI
arXiv:2610.13057v1 Announce Type: cross Abstract: Counterfactual text editing is the ability to revise a given text in light of a hypothetical change while preserving facts not causally affected by that change. When asked what would have…
-
Evolving Creatives with Clicks: CTR-Driven Advertising image Editing via Attribute Disentanglement and Preference Alignment
arXiv cs.AI
arXiv:2610.13066v1 Announce Type: cross Abstract: The optimization objective for e-commerce advertising creatives has shifted from aesthetic quality to online business metrics like click-through rate (CTR). Recent methods generate…
-
Tabular Foundation Models Are Effectively Shallow
arXiv cs.AI
arXiv:2610.13253v1 Announce Type: cross Abstract: In-context-learning Tabular Foundation Models (TFMs) are uniquely suited to scenarios with limited resources or privacy constraints, as they solve new tasks without additional training…
-
SmolWAM: A Latent World-Action Model for Robotic Manipulation on Edge Devices
arXiv cs.LG
arXiv:2610.13740v1 Announce Type: cross Abstract: World-action models (WAMs) predict a robot's actions together with the observations those actions will produce. However, the strongest models rely on video diffusion transformers that are…