Artificial Intelligence
Artificial Intelligence news — Monday 28 September 2026
150 stories from 11 sources, every one in English. Each links to its publisher.
-
Developing a new era... "President of the State-of-the-art"
AI Times (Korea)
Translated from Korean
-
Towards safety cases for frontier AI training
OpenAI Blog
Our early guidelines for safety cases in frontier AI training cover technical safeguards, operational practices, and investigating misalignment incidents
-
Offers 5.5, Self-Resume 1st place...'Incomplete RSI'In front of July
AI Times (Korea)
Translated from Korean
-
A successful completion of the 9-ruft hill amplitude calculation of the Clodelo theory
AI Times (Korea)
Translated from Korean
-
5 global leaderboard seats with Bead Ramp, AI
AI Times (Korea)
Translated from Korean
-
“Human Scientists vs AI” Study Interpretation
AI Times (Korea)
Translated from Korean
-
Jib, Claude’s 5 Supported ‘Pocketmon’ Qualification
AI Times (Korea)
Translated from Korean
-
"GPU line knows that the sand is only 18 tons"...Environ logo truck picking seafing
AI Times (Korea)
Translated from Korean
-
Requid AI Accelerates VLM Abstract 3x With ‘Extra Decoding’
AI Times (Korea)
Translated from Korean
-
When can we say AI made a scientific discovery?
MIT Technology Review Biotech
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here . Last Wednesday, Anthropic announced that earlier this year it had launched a molecular biology lab…
-
U.S., Russia, United Arab Emirates, United States, United Arab Emirates, United States
AI Times (Korea)
Translated from Korean
-
Mobility “LG·Duty·Duty Robot Custom OnDevice AI Semiconductor Development”
AI Times (Korea)
Translated from Korean
-
Microsoft’s Copilot redesign provides an experimental opportunity
AI Business
The revamp transforms Copilot from a simple AI assistant into a content workspace, providing knowledge workers with advanced developer tools.
-
Amoday, SNL Feng Shui appeared..."AI is devil, I'm the Creator"
AI Times (Korea)
Translated from Korean
-
Ukraine, "Robot Rover" rally.. .Dron Ears Ground Battleship Infinite Fire
AI Times (Korea)
Translated from Korean
-
Jeonnam-Gwangju, 443 billion won for the private business in 2027...
AI Times (Korea)
Translated from Korean
-
Guided Merge Sort : An Optimized Sorting that Picks the Best from Ordinary and Multi-Way Merge Sort Algorithms
Towards Data Science
A solution where the "goto" operator becomes irreplaceable. The post Guided Merge Sort : An Optimized Sorting that Picks the Best from Ordinary and Multi-Way Merge Sort Algorithms appeared first on Towards Data Science .
-
13.7 million visits per year
AI Times (Korea)
Translated from Korean
-
“Night Festa” will be held on October 3rd...Books·Music·Film Hall
AI Times (Korea)
Translated from Korean
-
Copyright (C) 2018. K-POP All Rights Reserved.
AI Times (Korea)
Translated from Korean
-
Bill Gates "The AI safety device rejection of Trumf.. . China will also be regulated"
AI Times (Korea)
Translated from Korean
-
Selection of 5 patents and patents
AI Times (Korea)
Translated from Korean
-
China ranked No. 1 in AI, ranked No. 77% in national residency
AI Times (Korea)
Translated from Korean
-
'Professor of the Government'...Hizhou Sang-won, CEO of OpenAI·&Roppy
AI Times (Korea)
Translated from Korean
-
How to Make Your Own JEV Model from an Open LLM
Towards Data Science
Turn a small open-source Qwen LLM into a fast, single-pass text classifier by swapping its language-modeling head The post How to Make Your Own JEV Model from an Open LLM appeared first on Towards Data Science .
-
OpenAI "ChatGPT iPhone Integrated Performance Reproduction"...Change Partnership Crack
AI Times (Korea)
Translated from Korean
-
AI safety is falling behind the pace of development
AI Business
Businesses need continuous monitoring, stronger guardrails and human oversight to keep increasingly capable AI systems under control.
-
OpenAI Agent, Un Data Blocker bypass..."Scanning Scan" capture
AI Times (Korea)
Translated from Korean
-
Makinavars, 'AI Truemo' military command control system applied.. .After evacuation
AI Times (Korea)
Translated from Korean
-
ScaberLab launches AI character chat ‘Jet’ new feature. . . . . . . .
AI Times (Korea)
Translated from Korean
-
POSGCUBE, building AI agent service to Hana Bank.. ."Finance AX Acceleration"
AI Times (Korea)
Translated from Korean
-
Kakao Co., Ltd. established the CEO of Kakao X in collaboration with Kakao Corporation
AI Times (Korea)
Translated from Korean
-
SKT·Samsung Electronics·Hanwha Financial Group established the first ‘5G Smart Office’
AI Times (Korea)
Translated from Korean
-
Import AI 474: Platonic mindspace; TPUs in space; Zhipu starts an outer RSI loop
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 Are minds patterns from a Platonic space, with bodies and…
-
The AI That Learned to Understand Long After It Stopped Trying
Towards Data Science
A small, strange discovery in machine learning called grokking The post The AI That Learned to Understand Long After It Stopped Trying appeared first on Towards Data Science .
-
The Download: rogue agent liability and the AI Hype Index
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. Who’s liable when AI agents go rogue? Over the past few months, a cascade of cyberattacks by AI…
-
Meta Launches ‘Horizontal Creaate’ for AI game with smartphone
AI Times (Korea)
Translated from Korean
-
'AI chip shortage' China, Allows to buy GPUs for NVIDIA PC
AI Times (Korea)
Translated from Korean
-
Antropic, Launched ‘Colroad Sonet 5.5’.. . OpenAI 'DEVDAY'
AI Times (Korea)
Translated from Korean
-
U.S.A., 2Q Solar DC 11.4GW Added...Up to 45%
AI Times (Korea)
Translated from Korean
-
Copyright (C) 2010 - 2018 SQUARE ENIX CO., LTD. All Rights Reserved.
AI Times (Korea)
Translated from Korean
-
How to Catch Data Drift When Every Feature Looks Normal
Towards Data Science
Detect hidden shifts in feature relationships with adversarial validation and scikit-learn The post How to Catch Data Drift When Every Feature Looks Normal appeared first on Towards Data Science .
-
Established 6MW of Tajkistan solar plant in Korea...
AI Times (Korea)
Translated from Korean
-
Holo4: powering generalist computer-use agents
Hugging Face Blog
-
[AIReader's Book] One Sweat My First AI Agent
AI Times (Korea)
Translated from Korean
-
Who’s liable when AI agents go rogue?
MIT Technology Review Biotech
MIT Technology Review Explains : Let our writers untangle the complex, messy world of technology to help you understand what’s coming next. You can read more from the series here . Over the past few months, a cascade of cyberattacks by AI…
-
The Lenfest Institute grows landmark program with expanded OpenAI support
OpenAI Blog
OpenAI is expanding the Lenfest AI Collaborative and Fellowship Program with $5 million in funding and up to $5 million in software credits and engineering support.
-
Trump, Amoday and the First Grand Prix
AI Times (Korea)
Translated from Korean
-
Scale-invariant Gaussian derivative residual networks
arXiv cs.CV
arXiv:2603.02843v2 Announce Type: replace Abstract: Generalisation across image scales remains a fundamental challenge for deep networks, which often fail to handle images at scales not seen during training (the out-of-distribution…
-
Diagnosing the Sources of Compositional Failure in Vision-Language Models: A Controlled Analysis
arXiv cs.CV
arXiv:2609.31456v1 Announce Type: new Abstract: Vision-language models (VLMs) often struggle with compositional reasoning tasks, but the reasons for this underperformance remain unclear. A common hypothesis is that models struggle to…
-
The Communication Map of a Transformer
arXiv cs.CL
arXiv:2608.22007v2 Announce Type: replace-cross Abstract: The components of a transformer communicate by writing to and reading from a shared residual stream, and the mechanistic interpretability literature has mapped these connections by…
-
Influence Diagnostics in High-dimensional M-estimation: Precise Asymptotics
arXiv cs.LG
arXiv:2607.09250v2 Announce Type: replace-cross Abstract: The impact of a given training point on a statistical model can be measured through its leave-one-out influence on the model parameters, which quantifies how its removal from the…
-
Learning from Next-Frame Prediction: Autoregressive Video Modeling Encodes Effective Representations
arXiv cs.CV
arXiv:2512.21004v2 Announce Type: replace Abstract: Recent advances in pretraining general foundation models have significantly improved performance across diverse downstream tasks. While autoregressive (AR) generative models like GPT have…
-
Explainable Predictive Condition-based Maintenance of Naval-Propulsion Systems using Fuzzy Logic
arXiv cs.LG
arXiv:2609.24250v2 Announce Type: replace Abstract: The shipping industry has a significant impact on the global economy, emphasizing the need for operational availability and safety through the use of effective maintenance techniques…
-
On associative neural networks for sparse patterns with huge capacities
arXiv cs.LG
arXiv:2603.26217v2 Announce Type: replace-cross Abstract: Generalized Hopfield models with higher-order or exponential interaction terms are known to have substantially larger storage capacities than the classical quadratic model. On the…
-
Generating Legal Commentaries from Case Databases via Retrieval, Clustering, and Generation
arXiv cs.CL
arXiv:2605.24534v2 Announce Type: replace Abstract: We present a fully automated pipeline that transforms large collections of court decisions into legal commentaries for statutes - without providing any handcrafted doctrinal framework…
-
FlatClip: A Geometry-Aware Surface-Level Baseline for fMRI Representation Learning
arXiv cs.CV
arXiv:2609.31204v1 Announce Type: cross Abstract: Recent fMRI foundation models differ substantially in the spatial scale at which they represent brain activity. ROI- and connectivity-based models are efficient but coarse, whereas…
-
The Right Information Extraction Pipeline Depends on the Document: Accuracy-Energy Trade-offs for Small, Local Models
arXiv cs.AI
arXiv:2609.31341v1 Announce Type: new Abstract: Whether an information extraction pipeline should process page images or parsed text depends on the document, and the answer flips across the layout spectrum. We study this trade-off under a…
-
Nice Fold or Hero Call: Learning Budget-Efficient Thinking under Policy-Dependent Solvability
arXiv cs.AI
arXiv:2605.11625v2 Announce Type: replace Abstract: Large reasoning models (LRMs) improve problem solving through extended reasoning, but often misallocate test-time compute. Existing efficiency methods reduce cost by compressing reasoning…
-
Do Neural Networks Preserve Case Structure? Case-Based Decomposition, Interpretation, and Decision Consistency
arXiv cs.AI
arXiv:2607.11347v2 Announce Type: replace Abstract: Neural networks increasingly inform consequential decisions, making their reliability increasingly important. Yet their internal mechanisms provide little evidence of whether decisions…
-
MVVBench: Benchmarking 4D Reasoning in Vision-Language Models
arXiv cs.AI
arXiv:2609.30952v1 Announce Type: cross Abstract: Multi-view video understanding requires integrating spatial and temporal evidence across multiple, often non-overlapping camera streams: tracking entities as they transition between…
-
Asking For An Old Friend: Diagnosing and Mitigating Temporal Failure Modes in LLM-based Statutory Question Answering
arXiv cs.CL
arXiv:2605.23497v2 Announce Type: replace Abstract: Large language models are increasingly used for legal research, yet their fixed training cutoffs and reliance on static parametric knowledge are at odds with the evolving nature of…
-
Enhancing Photogrammetric Digital Surface Models with Pretrained Diffusion Models and Multimodal Conditioning
arXiv cs.CV
arXiv:2609.31199v1 Announce Type: cross Abstract: Large-scale Digital Surface Models (DSMs) can be produced cost-effectively from satellite images via stereo-photogrammetry. However, the resulting 3D maps are often contaminated by noise…
-
A Vision-Language Foundation Model for Precise and Comprehensive Brain Tumor Diagnosis from Preoperative Multimodal Data
arXiv cs.AI
arXiv:2609.16597v4 Announce Type: replace-cross Abstract: We developed BrainVLM to classify all 12 World Health Organization (WHO) 2021 brain tumor types. BrainVLM integrates an uncertainty quantification strategy to indicate prediction…
-
Lifted Bellman Linear Programming for Offline Reinforcement Learning
arXiv cs.LG
arXiv:2609.24489v2 Announce Type: replace Abstract: Offline reinforcement learning (RL) typically trains a critic by minimizing a regression loss against bootstrapped value targets stabilized by target networks with exponential moving…
-
Regularizing modality contribution drift in multimodal continual learning
arXiv cs.LG
arXiv:2607.27260v2 Announce Type: replace Abstract: Multimodal continual learning (MMCL) aims to acquire new knowledge from multimodal data while retaining previously learned knowledge. Existing MMCL methods primarily mitigate forgetting…
-
Sorry Robot, Happy Human: Vision-Language Models Read Only One of Two Legible Typographic Layers
arXiv cs.AI
arXiv:2609.31403v1 Announce Type: cross Abstract: Vision-language models (VLMs), despite their success in optical character recognition (OCR) tasks, are vulnerable to typographic attacks and have a fragile structure for images with…
-
Persistent Negatives for Adversarial Black-Box On-Policy Distillation
arXiv cs.AI
arXiv:2609.30864v1 Announce Type: cross Abstract: Black-box On-Policy Distillation (OPD) seeks to improve a student from its own generations when the teacher provides sampled responses but not token probabilities. Adversarial distillation…
-
BAT-CLIP: Trimodal Alignment of Brain, Audio and Text
arXiv cs.LG
arXiv:2609.31180v1 Announce Type: cross Abstract: Decoding and interpreting naturalistic speech from the brain increasingly relies on alignment to pretrained speech and language representation spaces. However, current CLIP-style…
-
Affective Flow Language Model for Emotional Support Conversation
arXiv cs.CL
arXiv:2602.08826v3 Announce Type: replace Abstract: Large language models (LLMs) have advanced emotional support conversation, but existing alignment methods rely mainly on sparse preferences at the response level or outcomes at the…
-
DanLing NestedTensor: Composable Multi-Ragged Tensors for Deep Learning
arXiv cs.AI
arXiv:2609.30379v1 Announce Type: cross Abstract: Variable-size inputs are common in deep learning, but dense batching allocates a shared envelope and spends computation on padding. The cost multiplies across varying axes: an explicit pair…
-
ENAS: An Efficient Hardware-Aware Neural Architecture Search Framework for TinyML on Resource-Constrained Microcontrollers
arXiv cs.AI
arXiv:2609.30272v1 Announce Type: cross Abstract: We present \textbf{ENAS}, a hardware-aware Neural Architecture Search (NAS) framework that combines a static feasibility check, a cell-based search space supporting standard…
-
Why Alzheimer's Speech Screening Fails to Generalize: Bridging the Deployment Gap via Cross-Corpus Evidence Anchoring
arXiv cs.CL
arXiv:2609.31293v1 Announce Type: cross Abstract: Speech-based screening is a promising, non-invasive approach for detecting Alzheimer's disease and related cognitive risks. However, models trained on a single domain often generalize…
-
Analyzing and Mitigating Cost-Inefficient Behaviors in Coding Agents
arXiv cs.AI
arXiv:2609.30725v1 Announce Type: new Abstract: Although effective, coding agents often incur substantial monetary costs. Their recurring cost-inefficient behaviors remain underexplored. We conduct the first study of behavioral cost…
-
TempQ-Jail: Query-Constrained Candidate Ranking for Text-to-Video Jailbreak Attacks
arXiv cs.CV
arXiv:2609.31032v1 Announce Type: cross Abstract: Existing text-to-video (T2V) jailbreak methods mainly seek more effective or stealthier attack candidates. In guarded T2V systems, however, video generation and security evaluation are…
-
DyMD: Preserving Interaction Dynamics through Distribution Matching Distillation in Few-Step Video World Models
arXiv cs.CV
arXiv:2609.31349v1 Announce Type: new Abstract: Large video diffusion models offer expressive priors for embodied prediction and learning, yet their many-step sampling remains costly for interactive downstream use. Distribution Matching…
-
Muslim: A Deployed Arabic Voice AI Platform for Grounded Islamic Knowledge
arXiv cs.CL
arXiv:2609.31511v1 Announce Type: new Abstract: We present Muslim, a production Arabic voice AI platform serving grounded, sourced Islamic knowledge to real users. Beyond a real-time voice pipeline (NeMo Arabic ASR, an OpenAI-compatible…
-
Does Uniform Discrete Diffusion Need Time?
arXiv cs.LG
arXiv:2609.30977v1 Announce Type: new Abstract: Uniform discrete diffusion models (UDMs) commonly use explicit time conditioning, but we find that it can often be unnecessary in practice. In this paper, we first show that the…
-
GVCC: Zero-Shot Video Compression via Codebook-Driven Stochastic Rectified Flow
arXiv cs.AI
arXiv:2603.26571v4 Announce Type: replace-cross Abstract: At ultra-low bitrates, high-fidelity reconstruction requires sampling plausible videos from the posterior rather than regressing to oversmoothed conditional means. We propose…
-
Probing Stability-Plasticity Tradeoffs in Agent Memory through Cognitive Experimental Paradigms
arXiv cs.CL
arXiv:2609.30558v1 Announce Type: new Abstract: Agent memory systems are increasingly used to maintain long-term user preferences, task states and evolving facts, but current evaluations often collapse memory behavior into final-answer…
-
Recursive Self-Improvement via On-Policy Distillation for Reasoning
arXiv cs.CL
arXiv:2609.30652v1 Announce Type: new Abstract: On-policy distillation (OPD) trains a student model by having it generate trajectories, then matching its next-token predictions with an external teacher's next-token predictions. This…
-
Rufus-Air: An Open LLM Post-Training Recipe
arXiv cs.AI
arXiv:2609.29421v2 Announce Type: replace-cross Abstract: Rufus-Air is an open and reproducible post-training recipe on GLM-4.5-Air-Base (106B-A12B), organized as a serial pipeline of eight stages: SFT, Reasoning RL, Coding RL…
-
Modeling Student Sensemaking with LLMs and Knowledge-Graph-Guided Inference
arXiv cs.CL
arXiv:2609.31046v1 Announce Type: new Abstract: Collaborative science learning requires nuanced interpretation of student dialogue to characterize how learners identify knowledge gaps, build explanations, and work toward resolution - a…
-
Staged Depth Training: A Representation Curriculum for PINNs
arXiv cs.LG
arXiv:2609.30299v1 Announce Type: new Abstract: Representation quality is a central determinant of PINNs' performance, yet standard training leaves representations to emerge implicitly while fitting the final solution. We introduce…
-
A Framework for Identifying, Categorizing, and Explaining Bias in AI-Generated Code
arXiv cs.AI
arXiv:2609.30642v1 Announce Type: cross Abstract: As Large Language Models (LLMs) become integrated into software development workflows, concerns regarding unintentional biases in AI-generated code. Although evidence suggests these biases…
-
DecoVAE: a Lightweight Interpretable Trend-Seasonal VAE Framework for Efficient Probabilistic Time Series Forecasting
arXiv cs.LG
arXiv:2608.20052v3 Announce Type: replace Abstract: Probabilistic time series forecasting remains challenging, largely because modeling distinct trend and seasonal dynamics requires specialized approaches. Existing methods often fail to…
-
Provably Safe Sim-to-Real Transfer
arXiv cs.AI
arXiv:2609.01418v2 Announce Type: replace-cross Abstract: We address safe sim-to-real transfer, in which an agent leverages an imperfect simulator and limited real-world interaction while ensuring safety throughout data collection in the…
-
PIA: A Personal Intelligence Agent Turning Health Conversations into Records and Records into Understanding
arXiv cs.CL
arXiv:2609.31255v1 Announce Type: new Abstract: General-purpose agent memory summarizes conversations: it extracts salient snippets, embeds them, and retrieves the top-k into the prompt. A health agent cannot run on summaries: a dose…
-
Improving Molecular-Morphology Contrastive Pretraining using Deep-Learning-based Morphology Profiles
arXiv cs.LG
arXiv:2609.30433v1 Announce Type: new Abstract: Recent advancements in image-based profiling techniques have enabled the collection of high-volume cell morphology data, allowing new molecular embedding models to learn from the experimental…
-
The Residual Stream's Effective Depth
arXiv cs.LG
arXiv:2609.31098v1 Announce Type: new Abstract: We introduce \emph{effective depth} ($\Deff$), a scalar diagnostic that treats the layer-wise residual stream of a transformer as a discrete-time process, measures how representation…
-
Foundations of Reinforcement Learning and Interactive Decision Making
arXiv cs.LG
arXiv:2312.16730v2 Announce Type: replace Abstract: Interactive decision making is the problem of learning to act well in an unknown environment, using the data that one's own actions generate to continuously improve, and arises in…
-
Tabular Imbalanced Learning: A Survey, Benchmark, and Practical Guide
arXiv cs.LG
arXiv:2605.14915v2 Announce Type: replace Abstract: Imbalanced learning remains a fundamental challenge in tabular data applications. Despite decades of research and numerous proposed methods, there is still limited systematic…
-
PUBG Ally: A Conversational Embodied Agent as an AI Teammate
arXiv cs.CL
arXiv:2609.29837v2 Announce Type: replace-cross Abstract: We introduce PUBG Ally, an embodied agent for PUBG: BATTLEGROUNDS that can reason, act autonomously, and play alongside players as a voice-enabled teammate. Building such a teammate…
-
Semantic Navigation for Issue Localization in Code Repository
arXiv cs.AI
arXiv:2609.31176v1 Announce Type: new Abstract: Repository-level issue localization aims to identify and rank the files and functions relevant to resolving a reported issue. LLM agents approach this task iteratively: they identify a set of…
-
REALMS: An AI-Assistant Conversational System for Real-Time Exact Audience Sizing over High-Dimensional Nested Profiles
arXiv cs.CL
arXiv:2609.30547v1 Announce Type: new Abstract: Audience sizing is a critical component of digital marketing. It enables precise resource allocation, campaign planning, and performance optimization. Traditional approaches using skeleton…
-
Adaptive Random Matrices in Gaussian Bandits: Spectral Universality and Selection-Induced Outliers
arXiv cs.LG
arXiv:2609.30321v1 Announce Type: cross Abstract: Adaptive arm selection changes the distribution of the observations collected by a bandit algorithm, but it need not change their limiting empirical spectrum. We study Gaussian bandit…
-
Compress What You See, Not What You Say: Anchored Context Distillation for Latent-Observation Software Engineering Agents
arXiv cs.AI
arXiv:2609.31430v1 Announce Type: new Abstract: Tool observations dominate the context of software-engineering agents, making long interaction histories costly to maintain. Existing context compression methods can discard information…
-
EXAONE Demand 1.0: A Time Series Foundation Model for Demand Forecasting
arXiv cs.LG
arXiv:2609.30880v1 Announce Type: cross Abstract: Time series foundation models (TSFMs) are pretrained on series from diverse domains, where demand series make up only a small fraction. Demand data has properties that such corpora rarely…
-
Towards Unified Dynamic Face Landmark Detection
arXiv cs.CV
arXiv:2608.10346v2 Announce Type: replace Abstract: Although advancements in face landmark detection (FLD) methods continue to push performance boundaries, they overlook two major functional limitations: (1) different network parameters…
-
You've Seen Enough: Quality-Constrained Image Coding for Machines
arXiv cs.CV
arXiv:2609.25108v2 Announce Type: replace Abstract: Visual data is increasingly consumed by machine-vision systems rather than by human observers. Image Coding for Machines (ICM) compresses images assuming the main observer is a computer…
-
Causal Retention in Interactive Agents: Interface Factorization and Selective Adaptation
arXiv cs.LG
arXiv:2609.30650v1 Announce Type: new Abstract: Task performance need not determine which intervention mechanism an agent retains. We study causal retention: whether a frozen learned state answers a mechanism-probe map fixed independently…
-
Decodable In-Context State and Model Output Across Training
arXiv cs.LG
arXiv:2609.31401v1 Announce Type: new Abstract: Prior work established that a probe can decode an in-context binding on model errors and that probe-guided steering can repair some of them. We follow probe accuracy, model output, and…
-
SPADE: Escaping the Popularity-Similarity Frontier to Measure Serendipitous Recommendations
arXiv cs.LG
arXiv:2609.31164v1 Announce Type: cross Abstract: Recommender systems engineer serendipity to foster active exploration and break predictable consumption cycles. The problem with existing offline beyond-accuracy metrics is that they often…
-
Differentially-Private Decision Trees and Provable Robustness to Data Poisoning
arXiv cs.LG
arXiv:2305.15394v3 Announce Type: replace Abstract: Decision trees are interpretable models that are well-suited to non-linear learning problems. Much work has been done on extending decision tree learning algorithms with differential…
-
DyMD: Preserving Interaction Dynamics through Distribution Matching Distillation in Few-Step Video World Models
arXiv cs.AI
arXiv:2609.31349v1 Announce Type: cross Abstract: Large video diffusion models offer expressive priors for embodied prediction and learning, yet their many-step sampling remains costly for interactive downstream use. Distribution Matching…
-
Hide-and-Seek in Trajectories: Discovering Failure Signals for VLA Runtime Monitoring
arXiv cs.AI
arXiv:2605.30834v2 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models enable robots to follow natural language instructions and generalize across diverse tasks, but they remain vulnerable to execution failures that…
-
Quantum Diffusion Models for Medical Image Analysis
arXiv cs.AI
arXiv:2609.31070v1 Announce Type: cross Abstract: Quantum Machine Learning is a novel field of research aimed at devising machine learning approaches exploiting principles of quantum mechanics, such as superposition, entanglement and…
-
UQ-LOB: Uncertainty-Aware Limit Order Book Mid-Price Forecasting
arXiv cs.AI
arXiv:2609.31491v1 Announce Type: new Abstract: Forecasting short-horizon mid-price movements from limit order book (LOB) data is central to algorithmic trading, yet most deep LOB forecasters are point predictors: they output a direction…
-
Resource-Optimized and Energy-Aware Agentic AI Framework Anchored on Blockchain for Secure Software Supply Chains
arXiv cs.AI
arXiv:2609.31282v1 Announce Type: cross Abstract: This paper proposes a blockchain-backed agentic security framework designed to safeguard the complete software development lifecycle (SDLC) while also securing the agentic AI components…
-
Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework
arXiv cs.AI
arXiv:2609.30484v1 Announce Type: new Abstract: While large language models (LLMs) have achieved remarkable linguistic capabilities, a profound question lingers at their core: do these models truly comprehend context or simply excel at…
-
An End-to-End Pipeline for Causal ML with Continuous Treatments: An Application to Financial Decision Making
arXiv cs.LG
arXiv:2609.30396v1 Announce Type: cross Abstract: This paper presents an end-to-end causal machine learning (ML) pipeline designed for real-world applications with continuous treatments. The proposed framework consists of six sequential…
-
NEXT: Physics-Informed Neuro-Spectral Exponential Time Differencing Architectures
arXiv cs.LG
arXiv:2609.31539v1 Announce Type: new Abstract: Physics-Informed Neural Networks (PINNs) build neural representations of time-dependent PDE solutions, naturally incorporating physics knowledge and observational data, which makes them well…
-
CT-Merging: Consensus Directions and Task-Specific Scaling for LoRA Adapter Merging
arXiv cs.LG
arXiv:2607.20561v2 Announce Type: replace Abstract: LoRA merging methods increasingly operate on the low-rank structure of task updates, yet how the common subspace is estimated and how coefficients are assigned after recomposition are…
-
Conservation Buys Stability and Factoring Buys Counterfactuals in Physical World Models
arXiv cs.LG
arXiv:2609.19674v2 Announce Type: replace Abstract: A learned simulator can reproduce its training conditions accurately yet fail in two distinct ways once those conditions change. Over long rollouts, small errors accumulate until the…
-
Hill Sampling for Test-Time Scaling: A Simple and Better Alternative to Repeated Sampling, Evolution, and Training
arXiv cs.AI
arXiv:2609.25510v2 Announce Type: replace-cross Abstract: Large language models (LLMs) can improve solutions to verifiable scientific and algorithmic problems by spending additional computation at test time. Recent systems achieve strong…
-
DreamSat-Pose: Spacecraft Pose Estimation from Single-View 3D Reconstructions and Learned 2D-3D Feature Matching
arXiv cs.LG
arXiv:2607.13449v2 Announce Type: replace-cross Abstract: 6-DoF pose estimation is a critical task in autonomous rendezvous and proximity operations. In the case of an unknown target, this task becomes challenging as it shall be paired…
-
Highlight-Then-Summarize: Learning to Compress Evidence for Long-Context Understanding
arXiv cs.CL
arXiv:2609.31382v1 Announce Type: new Abstract: Long-context understanding requires large language models (LLMs) to reason over lengthy documents, conversations, and code, yet task-relevant evidence is often sparse and scattered amid…
-
Online Learning via Learned Latent Bayesian Tracking
arXiv cs.LG
arXiv:2609.31559v1 Announce Type: new Abstract: Online learning in non-stationary environments requires models to adapt rapidly from streaming data under strict computational constraints. A principled approach casts online learning as…
-
New LoRA Skills Should Read but Never Write
arXiv cs.LG
arXiv:2609.31600v1 Announce Type: new Abstract: Low-rank adapters (LoRA) make it cheap to fine-tune a large language model once per task, but combining several independently trained adapters into one model remains difficult: merging the…
-
Manifold Projection and Iterative Autoencoder Refinement for Masked Language Modeling
arXiv cs.LG
arXiv:2609.30288v1 Announce Type: cross Abstract: In Transformer-based masked language models, attention is the primary mechanism for context mixing, but there are other ways to mix data across tokens. Recent attention-free mixers replace…
-
StarWM: Self-Supervised Trained Attention Routing for Robust World Models
arXiv cs.LG
arXiv:2609.30667v1 Announce Type: cross Abstract: A robust world model must strike the balance between faithfully capturing environmental dynamics and abstracting away from irrelevant content. While reconstruction-based world models ensure…
-
Stochastic Bilevel Optimization with Heavy-Tailed Noise
arXiv cs.LG
arXiv:2509.14952v3 Announce Type: replace Abstract: This paper considers the smooth bilevel optimization in which the lower-level problem is strongly convex and the upper-level problem is possibly nonconvex. We focus on the stochastic…
-
DeepFedNAS: Efficient Hardware-Aware Architecture Adaptation for Heterogeneous IoT Federations via Pareto-Guided Supernet Training
arXiv cs.LG
arXiv:2601.15127v4 Announce Type: replace Abstract: Deploying federated learning across heterogeneous IoT device fleets requires tailored neural network architectures for each device class, yet existing Federated Neural Architecture Search…
-
What Do Tabular Foundation Models Compute In Context? In-Situ Representation Refinement through Attention-Gated Updates
arXiv cs.LG
arXiv:2609.27679v2 Announce Type: replace Abstract: A tabular foundation model must discover which distinctions matter for each new table without updating its parameters. We develop in-situ representation refinement: support labels guide…
-
NS-ATTENTION: Newton-Schulz Transformations of Attention Outputs in Vision Transformers
arXiv cs.LG
arXiv:2609.27735v3 Announce Type: replace Abstract: Newton-Schulz (NS) iteration has recently been used in the Muon optimizer to transform update matrices during the training of large language models. Motivated by its spectral effect, we…
-
Combee: Scaling Prompt Learning for Self-Improving Language Model Agents
arXiv cs.LG
arXiv:2604.04247v2 Announce Type: replace-cross Abstract: Recent advances in prompt learning allow large language model agents to acquire task-relevant knowledge from inference-time context without parameter changes. For example, existing…
-
MM-ContextFold: Context Folding for Multimodal Agentic Retrieval
arXiv cs.AI
arXiv:2609.23121v2 Announce Type: replace-cross Abstract: Multimodal Agentic Retrieval (MAR) requires agents to solve complex information-seeking tasks by iteratively invoking external tools. Typical frameworks such as ReAct maintain raw…
-
Below-ground Fungal Biodiversity Can be Monitored Using Self-Supervised Learning Satellite Features
arXiv cs.LG
arXiv:2604.09818v2 Announce Type: replace Abstract: Mycorrhizal fungi are vital to terrestrial ecosystem functioning. Yet monitoring their biodiversity at landscape scales is often unfeasible due to time and cost constraints. Current…
-
Too Sure to Be Safe: Model Calibration for Reliable Log Anomaly Detection
arXiv cs.AI
arXiv:2608.17965v4 Announce Type: replace-cross Abstract: Online log anomaly detection is critical for maintaining the reliability of large-scale computing systems. Although recent language model-based log anomaly detectors achieve strong…
-
How Far Can INRs Go? Cross-Domain Parameter-efficient INR-Based Semantic Segmentation for Brain MRI
arXiv cs.CV
arXiv:2609.31573v1 Announce Type: new Abstract: Biomedical image segmentation is central to medical image analysis, but practical deployment often faces limited annotations, memory constraints, and cross-site distribution shifts. Implicit…
-
MeshGraphNet-Transformer: Scalable Mesh-based Learned Simulation for Solid Mechanics
arXiv cs.LG
arXiv:2601.23177v4 Announce Type: replace Abstract: We present MeshGraphNet-Transformer (MGN-T), a novel architecture that combines the global modeling capabilities of Transformers with the geometric inductive bias of MeshGraphNets, while…
-
Combee: Scaling Prompt Learning for Self-Improving Language Model Agents
arXiv cs.AI
arXiv:2604.04247v2 Announce Type: replace Abstract: Recent advances in prompt learning allow large language model agents to acquire task-relevant knowledge from inference-time context without parameter changes. For example, existing…
-
LapDDPM: Spectral Perturbation Diffusion for Robust Single-Cell Manifold Generation
arXiv cs.AI
arXiv:2506.13344v2 Announce Type: replace-cross Abstract: Generating high-fidelity and biologically plausible synthetic single-cell RNA sequencing (scRNA-seq) data is a critical challenge in computational biology, driven by the need to…
-
FARE: Forensic Acceptance Region Estimation for Catching Bait-and-Switch Image Generators
arXiv cs.CV
arXiv:2609.30982v1 Announce Type: new Abstract: Modern AI image generators are increasingly deployed as opaque APIs, where customers can query the deployed service, but cannot inspect model weights or architecture. This creates a practical…
-
Open Vocabulary Domain Unlearning
arXiv cs.LG
arXiv:2609.31356v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) exhibit remarkable zero-shot generalization, yet they often encode unwanted or hazardous stylistic domains such as idealized textbook diagrams in medical AI or…
-
Geometric Inconsistency Localization in Multi-View Image Sets
arXiv cs.AI
arXiv:2609.31247v1 Announce Type: cross Abstract: Novel view synthesis (NVS) models can produce realistic new views of the same scene from different viewpoints. However, these generated views are not always geometrically consistent with…
-
Bootstrapping Conversational Recommendation Agents At Spotify: Synthetic Data Generation and Self-Improvement Loops
arXiv cs.AI
arXiv:2609.30297v1 Announce Type: cross Abstract: Conversational recommendation agents are a new paradigm for content discovery, enabling users to express complex intents through natural language (e.g., "recommend Italian indie artists I…
-
Common-Mode Collapse and Recovery in Direct Feedback Alignment
arXiv cs.LG
arXiv:2609.31589v1 Announce Type: new Abstract: Direct feedback alignment (DFA) trains hidden layers through fixed random projections of output error. With tanh hidden units and independent sigmoid outputs, plain stochastic gradient…
-
At FullTilt: Real-Time Open-Set 3D Macromolecule Detection Directly from Tilted 2D Projections
arXiv cs.CV
arXiv:2604.10766v5 Announce Type: replace Abstract: Open-set 3D macromolecule detection in cryogenic electron tomography eliminates the need for target-specific model retraining. However, strict VRAM constraints prohibit processing an…
-
CG-Probes: Recovering Guardrail Directions from Patient Query Embeddings
arXiv cs.CL
arXiv:2609.31062v1 Announce Type: new Abstract: Patient-facing AI assistants promise valuable support to patients, but incoming queries can pose medical risks. To create guardrails, we work with oncologists to define three ordinal risk…
-
Developing a Roadmap to an AI-first Organization: A Case Study in Embedded Software Development
arXiv cs.AI
arXiv:2609.30863v1 Announce Type: cross Abstract: The emergence of AI agents is expected to reshape software engineering by moving beyond AI as assistants towards systems capable of planning, executing, and evaluating development tasks…
-
Prompt Minimization: Reducing Input Redundancy Without Sacrificing Output Fidelity
arXiv cs.AI
arXiv:2609.31505v1 Announce Type: new Abstract: Despite the growing capabilities of large language models (LLMs), prompt design remains largely heuristic and ad hoc. This project will explore $\textit{prompt minimization}$, the process of…
-
Aim Short to Reach Far: Your Frozen World Model Can Plan Better Than You Think
arXiv cs.LG
arXiv:2609.30036v2 Announce Type: replace Abstract: Planners built on visual world models commonly score each predicted outcome by its distance to the encoded goal image. We show that this target can limit control even with exact dynamics…
-
Metacognitive Selective Ensemble for Mobile Systems
arXiv cs.LG
arXiv:2609.31031v1 Announce Type: new Abstract: Deep ensembles improve robustness in mobile sensing, but repeatedly executing many models over continuous sensor streams is costly. Selecting only a few members reduces this cost, yet…
-
GVCC: Zero-Shot Video Compression via Codebook-Driven Stochastic Rectified Flow
arXiv cs.CV
arXiv:2603.26571v4 Announce Type: replace Abstract: At ultra-low bitrates, high-fidelity reconstruction requires sampling plausible videos from the posterior rather than regressing to oversmoothed conditional means. We propose Generative…
-
A Multi-Stage Framework for Kuzushiji Character Recognition in Japanese Historical Documents
arXiv cs.CV
arXiv:2602.19086v3 Announce Type: replace Abstract: Kuzushiji was a widely used cursive writing system in pre-modern Japan. Due to simplification and glyph variation, most modern Japanese readers cannot read Kuzushiji characters…
-
Learning Hierarchical Causal Representations of the Effects of Forcings on Temperature in Climate Models
arXiv cs.LG
arXiv:2609.30995v1 Announce Type: new Abstract: Machine learning (ML) emulators provide a fast and cost-effective method to simulate climate change scenarios after being trained on Earth System Models projections. However, the black-box…
-
First-Order Stationarity of Reverse Diffusions
arXiv cs.LG
arXiv:2609.31612v1 Announce Type: cross Abstract: Recent literature has shown a strong connection between optimization and sampling. We develop the corresponding first-order theory for diffusion models. First, the SDE-based reverse-time…
-
Adapting Visualization Techniques for Time-Series Anomaly Detection: From Convolutional Neural Networks to Convolutional-Recurrent Neural Networks
arXiv cs.CV
arXiv:2411.04707v4 Announce Type: replace Abstract: Deep neural networks achieve strong performance on complex tasks but are often regarded as "black boxes," which limits their adoption in domains where transparency is essential. This lack…
-
LadderMIL: Multiple Instance Learning with Coarse-to-Fine Self-Distillation
arXiv cs.CV
arXiv:2502.02707v5 Announce Type: replace Abstract: Multiple Instance Learning (MIL) for whole slide image (WSI) analysis in computational pathology often neglects instance-level learning as supervision is typically provided only at the…