Artificial Intelligence
Artificial Intelligence news — Thursday 01 October 2026
150 stories from 11 sources, every one in English. Each links to its publisher.
-
Meta expands Muse to small businesses
AI Business
The move comes as agentic AI plays an increasingly big part in business operations, sparking oversight concerns.
-
Rable-up "AI intelligent measures the era... 1Agent infrastructure build"
AI Times (Korea)
Translated from Korean
-
"Notifications and Robot Controls"...Notifications for developers on the "Safety Device O-Track"
AI Times (Korea)
Translated from Korean
-
Nvidia announces ‘Fisys-Rank’, a video-producing framework for physical law
AI Times (Korea)
Translated from Korean
-
OpenClo opens an AI agent control platform 'OCE'
AI Times (Korea)
Translated from Korean
-
Inception, the voice agent ‘Mercurie Voice’ is released...
AI Times (Korea)
Translated from Korean
-
OpenAI "China Doorshot's 'Protected Abstract' Blocks large extraction attempts"
AI Times (Korea)
Translated from Korean
-
Linkedin "AI Safety Jobs 91% Revenue.. . Employment Hanpa's niche growth"
AI Times (Korea)
Translated from Korean
-
NAVER, web browser "AI chat" is installed..."Ta service introduction review"
AI Times (Korea)
Translated from Korean
-
AI price reduction, the highest speed in technology history...Agricultural sustainability is ‘unemployment’
AI Times (Korea)
Translated from Korean
-
The eternal complement
OpenAI Blog
Advanced AI may matter most for the routine work behind breakthrough ideas. Explore why execution could shape the next economy and the pace of progress.
-
Humanoid buildup overlooks the robots already running factories
AI Business
Millions of robotic machines are already operating in factories worldwide, but oversight and safety standards risk falling behind.
-
Achievement of 9000 billion investments from GMI and Nvidia
AI Times (Korea)
Translated from Korean
-
Copyright (C) 2017 SHINJUKU. All Rights Reserved.
AI Times (Korea)
Translated from Korean
-
Suncheon, Food & Art Festival Opening 16th... 10th Anniversary “Minari Green”
AI Times (Korea)
Translated from Korean
-
Copyright(c) 2014. All rights reserved.
AI Times (Korea)
Translated from Korean
-
"A AI company IPO must be blocked even when it falls"... Powerful warning of Microbrew
AI Times (Korea)
Translated from Korean
-
UK MI5, a Chinese academic penetration to ‘give alarm’...At the need to stop cooperation
AI Times (Korea)
Translated from Korean
-
Apple, 13th Smart Home Hub Revealed...'AI Shiri's Challenge in Amazon and Google
AI Times (Korea)
Translated from Korean
-
How Albertsons Companies is reimagining retail from the inside out
OpenAI Blog
Albertsons Cos. is using ChatGPT Enterprise and the OpenAI API to help teams work faster and make grocery shopping easier for millions of customers.
-
Meta, AI Data Center ‘Label Facility’ designated... $ billion tax-free public commentary
AI Times (Korea)
Translated from Korean
-
Autoencoders vs. PCA: I Rigged the Test and PCA Still Won
Towards Data Science
A theoretical advantage that didn't survive contact with a real benchmark. The post Autoencoders vs. PCA: I Rigged the Test and PCA Still Won appeared first on Towards Data Science .
-
U.S. solar module prices increase in commercial and large system prices
AI Times (Korea)
Translated from Korean
-
"Offline store analysis also AI"...Deeping soothing, "sai"
AI Times (Korea)
Translated from Korean
-
KT opens hospital AX professional internal organization... “Improving patient convenience and operation efficiency with medical specialty AI”
AI Times (Korea)
Translated from Korean
-
No.1 Global Benchmark for Non-Drawing and Legal Data Learning
AI Times (Korea)
Translated from Korean
-
Government Increases Power Supply Capacity 52GW...2030 Renewable Energy 100GW Response
AI Times (Korea)
Translated from Korean
-
OpenAI, 'GPT-Synopsis' development agreement for semiconductor design... 'Share profit' application
AI Times (Korea)
Translated from Korean
-
Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs
Hugging Face Blog
-
Can an Apartment Search Agent Call the Model Fewer Times and Still Find Good Matches?
Towards Data Science
I traced 2,500 listing checks with Weights & Biases Weave, removed avoidable model work one change at a time, and scored every version against the same answers. The post Can an Apartment Search Agent Call the Model Fewer Times and Still…
-
"Feed AI Agent" comment on...Mi FTC investigation·OpenAI first hydrogen
AI Times (Korea)
Translated from Korean
-
OpenAI Applys ‘Desicell API’ in its own agent... AI Architecture edition changed
AI Times (Korea)
Translated from Korean
-
Graph-centric agentic intelligence
Amazon Science Blog
Graphs encode relationships. In a world where AI agents need to reason about complex systems, not just retrieve information, graph structure provides the scaffolding for causal inference, dependency tracking, and compositional reasoning…
-
Copyright (c) 2014 YOKANAVI All Rights Reserved. Achievement of the Stock Market
AI Times (Korea)
Translated from Korean
-
What the ReLU Revolution Revealed About Biological Plausibility
Towards Data Science
The activation function was never a fixed biological commitment but a working hypothesis revised under empirical pressure. The post What the ReLU Revolution Revealed About Biological Plausibility appeared first on Towards Data Science .
-
Hyundai Motor Group, the national highway governmental AI and intelligent charging technology
AI Times (Korea)
Translated from Korean
-
"Photography Instant AI data conversion". . Korea Deep Learning, ‘Deep Agent’ Mobile Release
AI Times (Korea)
Translated from Korean
-
Meta, ‘Muse’, 29% of the monthly share price on the go
AI Times (Korea)
Translated from Korean
-
The Download: AI “mind-reading” and creative uses for small batteries
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. An AI “mind-reading” tool can reconstruct what you’re looking at based on a brain scan A new AI tool…
-
Google, Insider 'Jeminai 4 Argon' Performance Conference.. ."Venture Limits" Intellectual
AI Times (Korea)
Translated from Korean
-
Your AI Bill Is a Toll Booth. Stop Paying Twice.
Towards Data Science
The budget nobody saw coming The post Your AI Bill Is a Toll Booth. Stop Paying Twice. appeared first on Towards Data Science .
-
sLM 'Solar Mini 4' released... 'Pro·Mini' Turret
AI Times (Korea)
Translated from Korean
-
"Andropic body value must be equal to 1.5 trillion dollars"...A AI IPO "delivery" warning
AI Times (Korea)
Translated from Korean
-
An AI “mind-reading” tool can reconstruct what you’re looking at based on a brain scan
MIT Technology Review Biotech
A new AI tool can guess what you’re looking at just by analyzing your brain scans—and re-create that image with remarkable precision. It can go the other way, too, and predict a person’s brain activity based on what they’re looking at. In…
-
Dencom, an AI platform ‘Denthub 1.0’ released before dental care
AI Times (Korea)
Translated from Korean
-
How smaller, distributed batteries could help the grid
MIT Technology Review Biotech
If you want to install a big battery in New York City, you have to cut through a notoriously tough tangle of regulations. That’s made it difficult to get large energy storage projects on the grid. Faced with those obstacles, some startups…
-
[AIReader's Letter] Hegging Face Action
AI Times (Korea)
Translated from Korean
-
[Kyogo] AI version of the Big Macquarie's "AI River"
AI Times (Korea)
Translated from Korean
-
Google, ‘Jiminai 4 Argon’ released...In addition to the 1st episode, “General Use Limit”
AI Times (Korea)
Translated from Korean
-
Let the Carrier Carry the Attack: Preserving the Subject in Adversarial Image Generation
arXiv cs.AI
arXiv:2609.39723v1 Announce Type: cross Abstract: Strong unrestricted adversarial attacks can distort the primary object of an image, hereafter referred to as the subject. To preserve subject integrity without compromising attack…
-
ShieldCLIP: Selective Safety Alignment for Harmful Content Mitigation in Multimodal Foundation Models
arXiv cs.AI
arXiv:2609.39688v1 Announce Type: cross Abstract: Multimodal encoders such as CLIP underlie many downstream systems, but their web-scale training data embed harmful associations that safety alignment must suppress without unnecessarily…
-
A helps B while B hurts A: directed transfer in instruction-tuning mixture
arXiv cs.LG
arXiv:2609.39702v1 Announce Type: cross Abstract: Adapting a language model to a specialized corpus means choosing which instruction-tuning tasks to train on under a fixed budget, and testing one choice costs a fine-tuning run. Common…
-
Working Around the Compute Ceiling: Byte-Exact Memory in Galahad Makes LLM Reading a One-Time Cost LLM Reading a One-Time Cost
arXiv cs.AI
arXiv:2609.39358v1 Announce Type: cross Abstract: A transformer language model performs a bounded amount of computation per token, and recent work by Vishal Sikka, former CEO of Infosys, argues that this bound limits which tasks a model…
-
Patch Rebirth: Fast and Transferable Model Inversion of Vision Transformers
arXiv cs.CV
arXiv:2509.23235v3 Announce Type: replace Abstract: Model inversion is a widely adopted technique in data-free learning that reconstructs synthetic inputs from a pretrained model through iterative optimization, without access to original…
-
An Input-Frugal Deep Learning Framework for Weather-Driven National Crop-Yield Forecasting: A Case Study of Brazilian Soybean
arXiv cs.LG
arXiv:2609.38447v1 Announce Type: new Abstract: Reliable, timely crop-yield forecasts are essential for market stability and risk management, yet many approaches rely on costly or hard-to-scale inputs. We present a frugal, transferable…
-
FoCLIP: A Feature-Space Misalignment Framework for CLIP-Based Image Manipulation and Detection
arXiv cs.CV
arXiv:2511.06947v2 Announce Type: replace Abstract: The well-aligned attribute of CLIP-based models enables its effective application like CLIPscore as a widely adopted image quality assessment metric. However, such a CLIP-based metric is…
-
D-GAP: Improving Out-of-Domain Robustness via Dataset-Agnostic and Gradient-Guided Augmentation in Frequency and Pixel Spaces
arXiv cs.CV
arXiv:2511.11286v4 Announce Type: replace Abstract: Out-of-domain (OOD) robustness is challenging to achieve in real-world computer vision, especially in unsupervised domain adaptation scenarios, where shifts in image background, style…
-
Two-Step Data Augmentation for Masked Face Detection and Recognition: Turning Fake Masks to Real
arXiv cs.CV
arXiv:2512.15774v5 Announce Type: replace Abstract: The absence of large-scale masked face datasets challenges masked face detection and recognition. We propose a two-step generative data augmentation framework combining rule-based mask…
-
Cryo-Bench: Benchmarking Foundation Models for Cryosphere Mapping
arXiv cs.CV
arXiv:2603.01576v4 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…
-
EvoGuard: An Extensible Agentic RL-based Framework for Practical and Evolving AI-Generated Image Detection
arXiv cs.CV
arXiv:2603.17343v3 Announce Type: replace Abstract: The rapid proliferation of AI-Generated Images (AIGIs) poses severe misinformation risks, making AIGI detection critical yet challenging. Traditional detection paradigms mainly rely on…
-
Project and Mix: Task-Semantic Prototypes for Few-Shot Image Classification
arXiv cs.CV
arXiv:2603.24528v2 Announce Type: replace Abstract: Vision-language models like CLIP are trained with the objective of aligning text and image pairs. Beyond text prompts alone, recent works show that exploiting few-shot image embeddings…
-
One Adapter, Every Resolution: Gated Low-Rank Adaptation for Remote Sensing VLMs
arXiv cs.CV
arXiv:2605.07562v2 Announce Type: replace Abstract: Remote sensing imagery spans ground sampling distances (GSDs) from centimeters to tens of meters, so both the visual evidence for a geographic concept and the questions it can support…
-
scTrilemma: Balancing Identity, Invariance, and Fidelity in Single-Cell Representation Learning
arXiv cs.AI
arXiv:2609.38840v1 Announce Type: cross Abstract: Single-cell RNA-seq representation learning is fundamentally label-free: cell identities, states, and contexts are not fixed training targets, so what constitutes signal or nuisance is…
-
Rising Multi-Armed Bandits with Known Horizons
arXiv cs.LG
arXiv:2602.10727v3 Announce Type: replace Abstract: Rising Multi-Armed Bandits (RMABs) model sequential decision problems where each arm's expected reward improves with repeated pulls. In such problems, the value of investing in an arm…
-
Tissue Detection Determines False Positives in Diffusion-Based Histopathology Artifact Detection
arXiv cs.CV
arXiv:2609.40083v1 Announce Type: cross Abstract: One-class artifact detectors for whole-slide images learn normal tissue from a clean training pool and flag departures from it. The pool is built by a preprocessing pipeline whose…
-
Representation Risk in Pretrained Image Encoders
arXiv cs.CV
arXiv:2609.35470v1 Announce Type: cross Abstract: Applied researchers increasingly convert images into features with pretrained encoders, then use those features in a downstream prediction model. The encoder is often treated as an…
-
Examining Variation in How Guided AI Tutors Resolve Student Impasses
arXiv cs.AI
arXiv:2609.38346v1 Announce Type: new Abstract: When a student is stuck, a tutor faces the assistance dilemma: help given too early can hinder productive struggle, while help withheld too long leaves the student in a frustrating…
-
Less Data, Better Timing: Student-Curriculum Coupling for VLM On-Policy Distillation in Temporal Video Grounding
arXiv cs.CV
arXiv:2609.40055v1 Announce Type: new Abstract: On-policy distillation (OPD) provides dense supervision directly on student-generated trajectories, making it an effective post-training strategy for vision-language models in temporal video…
-
Grounding with Confidence: Controllable Generative Video Temporal Grounding
arXiv cs.CV
arXiv:2609.39883v1 Announce Type: new Abstract: Video temporal grounding supports applications such as video search, content review, and automated editing by localizing events described in natural language. Yet existing generative models…
-
When Instructions Retrieve Trajectories: Diagnosing and Mitigating Generalization Failures in VLA Models
arXiv cs.LG
arXiv:2609.39971v1 Announce Type: cross Abstract: Vision-language-action (VLA) models can exceed 90% success on in-distribution tasks and withstand nuisance changes that preserve the required action, yet fail under counterfactual changes…
-
Warm-starting PDE solvers with any-dimensional machine learning
arXiv cs.LG
arXiv:2609.38916v1 Announce Type: cross Abstract: Any-dimensional machine learning models, such as graph neural networks (GNNs), can be naturally trained and evaluated on inputs of different sizes and dimensions. Inspired by the GNN…
-
DyRAD: Radar Novel View Synthesis for Dynamic Driving Scenes
arXiv cs.CV
arXiv:2609.39841v1 Announce Type: new Abstract: Reconstructing dynamic driving scenes from recorded sensor data supports closed-loop evaluation of autonomous driving systems by synthesizing observations beyond the original trajectory…
-
P-SRM: Selective Recovery of Rejected Predictions in Visual Tracking
arXiv cs.CV
arXiv:2609.39832v1 Announce Type: new Abstract: Many visual tracking methods use rejection mechanisms to suppress unreliable predictions. However, these mechanisms can also reject correctly localized candidates, leaving useful information…
-
SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning
arXiv cs.AI
arXiv:2608.19842v3 Announce Type: replace Abstract: Agentic reinforcement learning (RL) has emerged as an important post-training approach for enhancing the capabilities of Large Language Models (LLMs). However, existing methods face a…
-
Diffusable Latents from Structure-Agnostic Distillation
arXiv cs.CV
arXiv:2609.39657v1 Announce Type: new Abstract: Distilling pretrained foundation models into an autoencoder bottleneck improves latent diffusability, enabling diffusion models to converge faster and reach higher sample quality. Standard…
-
ExpandDiff: Dynamic Range Expanding Diffusion for Single-Image HDR Reconstruction
arXiv cs.CV
arXiv:2609.39624v1 Announce Type: new Abstract: Single-image HDR reconstruction requires inferring missing detail while preserving the visible content of an LDR image. Differences in sensor dynamic range and exposure cause LDR images to…
-
Role-guided Speaker Deletion Verification in Clinical Psychiatry Speech Recordings with Audio Language Models
arXiv cs.LG
arXiv:2609.38491v1 Announce Type: new Abstract: Clinical research in psychiatry increasingly relies on large scale collection of spoken language data to identify acoustic and linguistic biomarkers. Yet evolving consent and protocol…
-
Unlocking the Critic: Reward-Free Policy Optimization for LLM Post-Training
arXiv cs.CL
arXiv:2609.37119v1 Announce Type: cross Abstract: Recent approaches to reinforcement learning (RL) post-training for large language models increasingly remove the critic to reduce training instability and memory overhead. Even where a…
-
Free Everywhere, Exact on Trees: PPO's Dropped Correction Buys Sample Efficiency Under Aggressive Reuse
arXiv cs.AI
arXiv:2609.39634v1 Announce Type: cross Abstract: Common policy improvement methods, including TRPO, PPO, and GRPO, estimate policy improvement under the behavioral policy's state-visitation distribution rather than the improved policy's…
-
Steepest Guidance: A Practical and Principled Approach to Inference-Time Alignment of Flow and Diffusion-based Models
arXiv cs.LG
arXiv:2609.39091v1 Announce Type: cross Abstract: Inference-time alignment of flow and diffusion-based models is critical for achieving flexible generative modeling. Theoretically, Doob's $h$-transform provides an elegant solution to this…
-
GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed
arXiv cs.CV
arXiv:2609.39600v1 Announce Type: new Abstract: Autoregressive (AR) grounding models serialize spatial predictions, introducing sequential latency and imposing a causal order on output tokens. We view grounding as visual evidence…
-
KilometerVision: A New Frontier for Large-Scale Spatial Intelligence in VLMs
arXiv cs.CV
arXiv:2609.39588v1 Announce Type: new Abstract: We push the frontier of large-scale spatial intelligence in Vision-Language Models (VLMs) and introduce the first benchmark that probes geographical layout understanding from real-world…
-
Zero-Compute Cross-Lingual Transferability Estimation Using Typological Feature Proxies
arXiv cs.AI
arXiv:2609.39640v1 Announce Type: cross Abstract: Cross-lingual transfer describes how knowledge in a source language benefits a target language. Measuring it quantitatively requires broad multilingual pre-training, as prior work has done…
-
Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs
arXiv cs.LG
arXiv:2609.32259v2 Announce Type: replace-cross Abstract: Recent multi-agent LLM systems increasingly combine heterogeneous models for specialized agent roles. However, text-based communication requires each receiver to prefill shared…
-
Evaluating and Benchmarking the System One Model Jev
arXiv cs.CL
arXiv:2609.37647v1 Announce Type: new Abstract: Jev is a commercial System One model from TypeSafe AI that does not generate text: given a state and typed questions, it returns a choice from fixed options, a position on a rubric, or the…
-
Cross-Linguistic Effects in Bilingual Phoneme BabyLMs
arXiv cs.CL
arXiv:2609.37121v1 Announce Type: new Abstract: Cross-linguistic effects are a central topic in bilingual first-language acquisition. Artificial learners can help investigate L1-L2 interactions by enabling controlled comparisons across…
-
PHASE: A Physiology-Guided Hierarchical Foundation Model for Intracranial EEG
arXiv cs.LG
arXiv:2609.36087v2 Announce Type: replace Abstract: Clinicians and neuroscientists have long analyzed intracranial electroencephalography (iEEG) through directly measurable physiological characteristics, which carry much of the information…
-
ECHO-G: Embodied Co-speech Humanoid mOtion Generation
arXiv cs.AI
arXiv:2609.39575v1 Announce Type: cross Abstract: Generating full-body co-speech motion for humanoid robots requires coordinating speech prosody, linguistic content, and embodiment-specific motion. To this end, we present ECHO-G, a…
-
Which papyrus HTR is good enough? Character-error-rate tolerance of four papyrological tasks on Greek texts
arXiv cs.CL
arXiv:2609.37755v1 Announce Type: new Abstract: Purpose: Most Greek papyri remain unpublished and undigitised; a handwritten text recognition (HTR) pipeline that transcribes them automatically would let scholars discover documents and…
-
Auditable Long-Term Memory: A Deterministic Retrieval Chain Measured at 479/475 of 500 on LongMemEval-S
arXiv cs.CL
arXiv:2609.38021v1 Announce Type: new Abstract: We evaluate an auditable long-term memory system on LongMemEval-S. Its retrieval chain uses hybrid candidate retrieval, cross-encoder reranking, coverage-first packet compilation, and…
-
Predicting Team Performance from Communications in Simulated Search-and-Rescue
arXiv cs.CL
arXiv:2503.03791v1 Announce Type: cross Abstract: Understanding how individual traits influence team performance is valuable, but these traits are not always directly observable. Prior research has inferred traits like trust from…
-
Group-Marginalized Self-Rewarding RL Drives Zero-Label Self-Evolving
arXiv cs.CL
arXiv:2609.36750v1 Announce Type: cross Abstract: Self-rewarding reinforcement learning (RL) enables large language models (LLMs) to self-evolve without human labels. Existing ensemble-based methods construct reward references from rollout…
-
Multimodal Detection of Higher-Order Behavioral Constructs: Self-Compassion in Structured Reflective Interaction
arXiv cs.CL
arXiv:2609.37148v1 Announce Type: cross Abstract: Many of the qualities that matter most in how people learn and grow, how someone regulates their emotions, reflects on a setback, or stays aware of others during a difficult conversation…
-
Port-Hamiltonian Latent Deliberation: Mitigating the Deliberation Drift Cliff in Test-Time Compute Scaling
arXiv cs.CL
arXiv:2609.37351v1 Announce Type: cross Abstract: Test-time compute scaling has emerged as a cornerstone of advanced machine reasoning, yet performing iterative deliberation directly within continuous latent representation spaces reveals a…
-
MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment
arXiv cs.CL
arXiv:2609.37574v2 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used to enrich user queries in information retrieval (IR) so that a standard retriever such as BM25 can bridge vocabulary gaps with the target…
-
Does Text Steer Neural PDE Surrogates? A Controlled Diagnostic with OperatorCLIP
arXiv cs.LG
arXiv:2609.38517v1 Announce Type: new Abstract: Lower error from a text-conditioned neural surrogate does not, by itself, show that the model uses the meaning of the text. We examine this attribution problem with OperatorCLIP, comparing an…
-
ORCA-bench: How Ready Are Language Model Agents for Oncall?
arXiv cs.CL
arXiv:2607.28545v3 Announce Type: replace Abstract: Large language models can write, patch, and search code, but oncall root cause analysis (RCA) demands something different: reasoning over noisy metrics, logs, traces, and source code…
-
OpenART: Scaling Agent Red Teaming via Open-Ended Environment Evolution
arXiv cs.CL
arXiv:2608.00677v2 Announce Type: replace Abstract: AI agents operate in persistent environments where early state changes can influence decisions far into the future. Unlike conventional language-model interactions, agent behavior is…
-
Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus
arXiv cs.CL
arXiv:2608.12149v3 Announce Type: replace Abstract: We present the first systematic study of massive activations (MAs) in layer-interleaved Hybrid linear attention large language models (HLA LLMs), examining their architectural…
-
The Copy Ceiling: An Input-Exposure Control for Ontology-Grounded Generation over Curated Corpora
arXiv cs.CL
arXiv:2609.24885v2 Announce Type: replace Abstract: We built a node that grounds a replaceable language model in a maintained ontology corpus, then asked what its successful-looking evaluation could support. Across ten models, grounding…
-
Isolated Sign Language Recognition for Icelandic Sign Language: Experiments in a Low-resource Setting
arXiv cs.CL
arXiv:2609.25862v2 Announce Type: replace Abstract: We present the first experiments on isolated sign language recognition (ISLR) for Icelandic Sign Language (\'ITM). We use \'ITM SignWiki, a dataset derived from a bilingual…
-
From Pixels to Pairs: A Comprehensive Benchmark of LLM-Driven Key-Value Extraction in Noisy Document Settings
arXiv cs.CL
arXiv:2609.17538v2 Announce Type: replace Abstract: Large language models (LLMs) have demonstrated strong capabilities in document key-value pair (KVP) extraction, yet controlled evaluations of their robustness to optical character…
-
Revisiting the Capacity Gap in Chain-of-Thought Distillation from a Practical Perspective
arXiv cs.CL
arXiv:2604.08880v2 Announce Type: replace-cross Abstract: Chain-of-thought (CoT) distillation transfers reasoning behaviors from a strong teacher to a smaller student, but prior work reports a capacity gap: distillation may fail when the…
-
Remember Your Trace: Memory-Guided Long-Horizon Agentic Framework for Consistent and Hierarchical Repository-Level Code Documentation
arXiv cs.CL
arXiv:2605.14563v3 Announce Type: replace-cross Abstract: Automated code documentation is essential for modern software development, providing the contextual grounding that both human developers and coding agents rely on to navigate large…
-
From Answers to Policies: Efficient In-Context Learning System through Emulating Expert Investigation
arXiv cs.CL
arXiv:2608.16831v2 Announce Type: replace-cross Abstract: Pretrained large language models offer a practical foundation for learning useful behavior from few task-specific examples. We argue that current prompt and context optimization…
-
RAISE: Diagnosing Acquisition Collapse in Costly LLM Signals
arXiv cs.CL
arXiv:2608.10441v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used as costly, on-demand components in real systems, but calling them indiscriminately can waste substantial compute, latency, and…
-
Decodable but Misrouted: Sparse Features Uncover a Readout Gap in Vision-Language Models for Harmful Meme Detection
arXiv cs.CL
arXiv:2609.18860v2 Announce Type: replace-cross Abstract: When large vision-language models misclassify harmful memes, the failure may reflect missing internal evidence or an inability to route represented evidence to their outputs. We…
-
Travel Time Prediction in Supply Chain Management Using Machine Learning
arXiv cs.LG
arXiv:2609.38190v1 Announce Type: new Abstract: The purpose of this research is to find data and methods using machine learning and deep learning to correctly predict the estimated travel time for transportation and logistics in a supply…
-
MILO: Automated Harness Discovery via Orchestrated Multi-Agent Evolution
arXiv cs.LG
arXiv:2609.38349v1 Announce Type: new Abstract: Modern agentic systems combine an AI model with a harness that controls execution and environmental interactions. Harness design strongly affects long-horizon performance, yet its…
-
Graph Anomaly Detection as Finite-Horizon Control: Training-Free Scoring via Empirical Bayes
arXiv cs.LG
arXiv:2609.38424v1 Announce Type: new Abstract: Node-level graph anomaly detection (GAD) identifies nodes whose attributes and interactions deviate from dominant graph regularities. Existing GAD models encode normality and anomaly scoring…
-
TTMark: Pairwise Distortion-Free Watermarking Beyond Single-Token Entropy
arXiv cs.CL
arXiv:2609.36372v1 Announce Type: cross Abstract: Distortion-free watermarking enables reliable attribution of machine-generated text while preserving output distribution. However, existing methods operate independently on each generated…
-
Vision Wormhole: Latent-Space Communication in Heterogeneous Multi-Agent Systems
arXiv cs.CL
arXiv:2602.15382v3 Announce Type: replace Abstract: Heterogeneous multi-agent systems combine models with different capabilities through a common communication interface. Exchanging internal states directly requires translating between…
-
SEA-CLIP-Tiny: Efficient Multilingual Text-Vision Embedding for Southeast Asian Languages
arXiv cs.CL
arXiv:2609.30739v2 Announce Type: replace Abstract: Multilingual text-vision embedding models are essential for cross-lingual image-text retrieval, but Southeast Asian languages remain poorly supported due to the region's linguistic…
-
ReLaG: A Scalable Framework Generalizing Random Splits to Data with Latent Relations
arXiv cs.LG
arXiv:2609.38538v1 Announce Type: new Abstract: Random splitting can yield non-independent train--test subsets when a dataset contains related samples, as is common in certain applications such as biochemical studies. This leads to overly…
-
Understanding Off- vs On-Policy Distillation: A Tale of Distinct Training Objectives
arXiv cs.LG
arXiv:2609.38666v1 Announce Type: new Abstract: On-policy distillation (OPD) learns from teacher feedback on student-generated responses and has shown promise in reducing forgetting relative to supervised fine-tuning (SFT). However, its…
-
VERA: Verifiable Feasibility Representations with Counterfactual Credit for Constrained Multi-Agent Control
arXiv cs.LG
arXiv:2609.38889v1 Announce Type: new Abstract: Constrained multi-agent control requires more than predicting rewarding actions: an action can cease to be executable as contact windows, shared capacity, and deadlines change. We introduce…
-
Unmerge: Efficient Machine Unlearning via Task Arithmetic
arXiv cs.LG
arXiv:2609.38895v1 Announce Type: new Abstract: Approximate machine unlearning seeks to remove the influence of a forget set from a trained model without full retraining. Existing gradient-based methods require data-dependent…
-
Certified Approximation for Interpretable Representer Landmarks
arXiv cs.LG
arXiv:2609.38901v1 Announce Type: new Abstract: Representer explanations rank the training landmarks that most influence a self-supervised representation. At scale, this ranking rests on up to four stacked approximations of the empirical…
-
Hard-Gate Candidacy in a Deployed Validator Suite
arXiv cs.LG
arXiv:2609.39037v1 Announce Type: new Abstract: Before a validator can be promoted to a hard gate on a deployment pipeline, it has to be shown that its firing separates outputs that reach users in working order from those that do not. We…
-
Shared Phase and Retention Control for Efficient Adaptive Spectral Recurrence
arXiv cs.LG
arXiv:2609.39082v1 Announce Type: new Abstract: As new evidence arrives, a sequence model must update what it remembers and how memory influences predictions. While Transformers incur computation and cache costs scaling with context…
-
A Generalisation Signal Need Not Be a Model-Selection Signal
arXiv cs.LG
arXiv:2609.39099v1 Announce Type: new Abstract: Model selection in computational biology often relies on validation data drawn from the training regime, even when deployment lies outside it. When validation no longer preserves which model…
-
Sharp Stationary Gaussian Approximation for Constant-Stepsize SGD
arXiv cs.LG
arXiv:2609.39144v1 Announce Type: new Abstract: We prove a sharp Gaussian approximation for the invariant law of constant-stepsize SGD with bounded additive noise generated by an exogenous uniformly ergodic Markov chain. For a smooth…
-
Right Answer, Wrong Mechanism: Detecting Pernicious Divergence in Causal Interventions
arXiv cs.LG
arXiv:2609.39243v1 Announce Type: new Abstract: Causal interventions such as activation patching and distributed alignment search (DAS) are the main tool for making mechanistic claims about neural networks. Recent work showed that these…
-
Jacobian Rank Collapse in Decision-Focused Learning
arXiv cs.LG
arXiv:2609.39261v1 Announce Type: new Abstract: Decision-focused learning (DFL) trains predictors through downstream objectives, but a different loss need not provide an independent parameter-update direction. We characterize this…
-
A Tilted Bowl Is Not a Slippery Slope: Compressing Looped Models
arXiv cs.LG
arXiv:2609.39277v1 Announce Type: new Abstract: Looped models reason by applying the same block of weights many times, so compressing that block saves memory traffic on every loop. Compressed looped models, however, often collapse, and the…
-
How Many Samples Are Enough for Learning Across Domains?
arXiv cs.LG
arXiv:2609.39336v1 Announce Type: new Abstract: Understanding the fundamental mechanisms of learning is essential for designing systems with strong generalization. Recent studies have shown that increasing the number of training domains…
-
SRJudge: Empowering Large Language Models with Selective Reasoning for Fine-Grained Knowledge Concept Tagging
arXiv cs.CL
arXiv:2609.36982v1 Announce Type: new Abstract: Knowledge concept tagging aims to assign specific concept or topic labels to educational content, which is essential for both educators and learners in traditional and online teaching…
-
Distilling What Matters: Confidence-Aware Selective Distillation for Large Language Models
arXiv cs.CL
arXiv:2609.36734v1 Announce Type: new Abstract: Knowledge Distillation (KD) trains a smaller-capacity student model to imitate a larger-capacity teacher model by matching output distributions, implicitly assuming the teacher to be a…
-
From Wrecks to Wisdom: Recovering Crash Mechanics from Real-World Multi-View Photos
arXiv cs.CV
arXiv:2609.39486v1 Announce Type: new Abstract: Estimating accident mechanics from real-world crashes is important for vehicle-safety analysis, injury modeling, crash-severity prediction, and operational workflows such as insurance claim…
-
Video2Skill: From Streaming Experience to Reusable Embodied Skills
arXiv cs.CL
arXiv:2609.36691v1 Announce Type: new Abstract: Manipulation behaviors vary widely across objects and scenes, but they share a small set of reusable skills, and planning with these skills helps embodied agents generalize to new tasks. Yet…
-
Similar Choices, Different Attention: Cross-Modal Associations in Humans and Vision-Language Models
arXiv cs.CL
arXiv:2609.36475v1 Announce Type: new Abstract: Cross-modal associations are systematic pairings of features across modalities, such as the association of 'bouba' with round shapes and 'kiki' with sharp shapes. Prior work has compared…
-
CAST: Causal Advantage-Structured Training with Spatially Grounded Compositional Rewards for Diffusion Models
arXiv cs.CV
arXiv:2609.39441v1 Announce Type: new Abstract: Online reinforcement learning has been extended to flow matching for diffusion model (DM) image generation. However, this paradigm faces three limitations: (1) Window selection. Existing…
-
Raw Imagery Impacting Your AI: Should You Care?
arXiv cs.CV
arXiv:2609.38265v1 Announce Type: cross Abstract: Onboard AI is gaining interest for space applications such as vessel, wildfire, and cloud detection, where real-time processing can improve mission reactivity and reduce downlink needs…
-
Resolving the Missing Financial Data Crisis: A Generative AI Pipeline for SEC 10-K Extraction
arXiv cs.CL
arXiv:2609.35864v1 Announce Type: new Abstract: SEC 10-K filings contain substantial financial information that is not consistently captured in structured datasets, creating a missing-data problem affecting over 70% of firms and half of…
-
PACT: Pairwise-Anchored Calibrated Tuning for Single-Token Typed Decisions
arXiv cs.CL
arXiv:2609.35865v1 Announce Type: new Abstract: Single-token typed-decision models answer a schema question by reading the logits of a few one-letter answer codes at a single position: they are fast and return a probability for every…
-
COBICount: Separating Object and Background Responses for Remote Sensing Object Counting Without Training on Target Data
arXiv cs.CV
arXiv:2609.39366v1 Announce Type: new Abstract: Remote sensing object counting estimates how many buildings, vehicles, or ships appear in overhead images. Most supervised counters predict a density map, whose sum gives the object count…
-
Hyperspherical Semantic Trajectory Analysis: Mapping Technological Diffusion across Academic Preprints, Patent Signals, and Compute Scaling
arXiv cs.CL
arXiv:2609.35845v1 Announce Type: new Abstract: Macroeconomic productivity metrics, such as Total Factor Productivity, register technological breakthroughs with multi-year reporting lags due to administrative survey intervals and national…
-
Can We Still Trust Disaster Social Sensing? Empirical Evidence on Detecting AI-Generated Social Media Posts
arXiv cs.CL
arXiv:2609.35821v1 Announce Type: new Abstract: Disaster social sensing converts public social-media posts into evidence for situational awareness and humanitarian needs, but generative artificial intelligence (AI) can produce plausible…
-
TexTailor: Texture-Preserving Video Virtual Try-On via Adaptive Garment Conditioning
arXiv cs.CV
arXiv:2609.39335v1 Announce Type: new Abstract: Video virtual try-on has attracted increasing attention due to its broad potential in digital fashion and intelligent e-commerce. However, existing methods primarily focus on low-resolution…
-
BadAction: Backdoor Attacks on Interactive Video Generation via Action-Guided Triggers
arXiv cs.CV
arXiv:2609.39047v1 Announce Type: new Abstract: Interactive video generation (IVG) models have achieved remarkable progress in producing controllable visual content guided by user-defined actions, yet their security vulnerabilities remain…
-
Disagreement-Regularized Imitation Learning for Image-Based Continuous Control with Gaussian and Beta Policies
arXiv cs.LG
arXiv:2609.38407v1 Announce Type: new Abstract: Purpose: Behavior cloning can accumulate errors when a learned controller visits states outside the demonstrated distribution. This study evaluates whether Disagreement-Regularized Imitation…
-
Learning Under Forgetting: Statistical Support-Selective Retention in Stochastic Training Dynamics
arXiv cs.AI
arXiv:2609.38768v1 Announce Type: cross Abstract: Prior work has shown that neural networks exhibit implicit biases toward low-complexity structure (e.g., spectral bias), memorization dynamics, and compression-like effects during training…
-
How to Run Statistics over LLM Judges and Trust the Results: Calibrated Inference for Small-Sample AI Evaluation with evalstats
arXiv cs.CL
arXiv:2609.35815v1 Announce Type: new Abstract: Researchers across academia increasingly base significance claims on LLM judge scores and small-sample AI evaluations. Yet without well-calibrated confidence intervals (CIs), hypothesis…
-
Lookahead-R: Budget-Aware Tool Retrieval via Execution-Centric Planning
arXiv cs.CL
arXiv:2609.35811v1 Announce Type: new Abstract: Tool retrieval is a critical bottleneck for LLM-based agents operating over large, heterogeneous API ecosystems. Existing approaches face an inherent trade-off: semantic retrievers are fast…
-
DC-SAE: Deep Compression Semantic Autoencoder for Faster Diffusion Convergence
arXiv cs.CV
arXiv:2609.39222v1 Announce Type: new Abstract: High-compression tokenizers are essential for scaling latent image generative models. However, aggressive compression creates a fundamental tradeoff between reconstruction fidelity and…
-
Can Multimodal Large Language Models Generate and Detect Multimodal Social Media Fake News?
arXiv cs.CL
arXiv:2609.35809v1 Announce Type: new Abstract: The rapid advancement of generative AI raises concerns about the misuse of Multimodal LLMs (MLLMs) for large-scale disinformation campaigns on social media. Despite existing research on…
-
FD-VAD: Semantic Endpoint Detection for Streaming Full-Duplex Speech
arXiv cs.CL
arXiv:2609.35791v1 Announce Type: new Abstract: Natural turn-taking in full-duplex voice interaction requires determining from partial speech whether a pause reflects hesitation or a completed conversational intent. Acoustic voice activity…
-
Uruqi: Learning Spatial Cognition from Visual Experience
arXiv cs.CV
arXiv:2609.39195v1 Announce Type: new Abstract: Spatial intelligence requires maintaining a coherent understanding of the world as the embodied agent moves. Like humans, the agent must use its own motion to interpret changes across…
-
The Golden Path Hypothesis: Reusable Schedules in Diffusion Caching
arXiv cs.LG
arXiv:2609.39343v1 Announce Type: cross Abstract: Diffusion caching accelerates generation by replacing transformer computation with cached or predicted features at selected denoising steps. We introduce the Golden Path Hypothesis (GPH)…
-
Targeted Retrieval, Compact Representations: How CoT Reasoning Improves Long-Context Counting
arXiv cs.LG
arXiv:2609.38958v1 Announce Type: cross Abstract: Large language models (LLMs) have been rapidly improving in long-context tasks, powered by Chain-of-Thought (CoT) reasoning. However, the internal mechanisms underlying this improvement…