Artificial Intelligence
Artificial Intelligence news — Friday 06 November 2026
150 stories from 7 sources, every one in English. Each links to its publisher.
-
[Nov. 6] "AI's AI Safety Deceased to Work Chucks"...Not Causes
AI has been a research to increase the safety of other AIs, and we have found that some actions are not properly performed and not knowing that. We use cookies to give you the best experience on our website. If you continue to use this…
-
Prompt: AI's next bottleneck is outside the data center
AI companies are racing to build data centers, but power shortages, regulatory hurdles and local opposition are slowing expansion.
-
How to Measure AI Adoption, Performance, and ROI With 12 KPIs
Includes a worked example with 50 users and $110,000 in first-year cost, plus baselines, targets, and ownership rules for every AI initiative. The post How to Measure AI Adoption, Performance, and ROI With 12 KPIs appeared first on Towards…
-
How to Manage Python Environments with uv Instead of venv and pip
A side-by-side walkthrough of venv plus pip and uv init, uv add, uv run, and uv sync, with a lockfile that recreates the project on another machine. The post How to Manage Python Environments with uv Instead of venv and pip appeared first…
-
Google's AI-first Search creates challenges for schools and businesses
Google is integrating AI deeper into Search, creating challenges for schools, advertisers and publishers, and raising web traffic issues.
- ToyBlocks: A controlled world for robot learning and evaluation
-
Build Your First Multi-Tool AI Agent with Ollama and Python
AI Agents for Beginners, Part 2: The agent picks between three database tools, uses each result to choose its next call, and stops at an eight-turn circuit-breaker if it gets stuck. The post Build Your First Multi-Tool AI Agent with Ollama…
-
The Download: a humanoid cleans house and Samsung builds for the AI boom
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. I watched a humanoid robot clean my apartment. It felt like a taste of the future. —Grace Huckins…
-
Give Your Python AI Agent Persistent Memory with One JSON File
Three Python classes, an OpenRouter key, and gpt-4.1-mini give the agent a calculator, a stub web search, and conversation recall that survives restarts. The post Give Your Python AI Agent Persistent Memory with One JSON File appeared…
-
I watched a humanoid robot clean my apartment. It felt like a taste of the future.
I’m not the sort of person who typically pays for a house cleaner. I’m too frugal, and I kind of enjoy putting on a podcast and going to town on my tub with a Magic Eraser. But when I saw an X post about a new humanoid-robot cleaning…
-
Google’s researcher, 8x big AI coding performance with 65B model...development cost 1/10
Resive Intelligence (Ricursive Intelligence), founded by Google’s researcher, revealed an open-source model before coding performance evaluation of 8x larger AI models. The cost of development than the existing leading model reveals that…
-
[Bulletin] Real World, Robot Foundation Model ‘Real Code’ demonstration
■ Real World (CEO Ryujung Hee) announced that the robot foundation model (RFM) ‘Redex’ will be held in Shanghai, China. Learning as a real data on Korean and Japanese manufacturing sites, supporting precision manipulation of five finger…
-
Kakao announces ‘Kanana-2-v-D Seaside’
In the ‘Solo Deside’ of TypeSafe AI and ‘Solo Deside’, Kakao also has a decision model for AI agent. Kakao (CEO Myeong-ah) announced that it has released a multimodal model ‘Kanana-2-v-side’ for AI decision-making support as an open source…
-
Falcon, Shoes manufacturer Changshin...AI
Palrantia Technologies (PLTR) announced that it has signed a partnership that applies ‘Foundry’ and AI platform ‘AIP’ across the development and production operation of the global shoe manufacturing company. In this collaboration, we…
-
nVidia promotes transition to 130 AI data centers in the US Post Office
Nvidia is promoting the implementation of a decentralized network that has been transformed into an AI data center for the US Post Office Building. We use cookies to ensure that we give you the best experience on our website. If you…
-
Antarctica, 'Call' game reproduction limit relief... "You can create a similar game"
Andropic has mitigated the game reproduction of the "Call Code" forbidden action. It’s time to recall the existing game, but it’s time to see specific games, and the work that creates a similar genre’s original game was allowed. Antropic…
-
France Solar New Connection 4.18GW...Capacity 33% reduction, Roof type 14% increase
This year of France’s 1-September, the solar new system capacity was about 4.18 gwat (GW), but the smallest facility in the residential area was reduced by 33.3%. During the same period, the roof-type solar was increased by 13.6% and the…
-
Antropic, Add Personal Financial Functions to Clad... Analyze Spending Accounts
Antropic has added ‘Personal Finance’ function that can analyze personal financial status by connecting bank accounts and credit cards to ‘Call’. We use cookies to ensure that we give you the best experience on our website. If you continue…
-
Japan and business photovoltaic panel 50tons and more will be promoted to reporting
Japan government promotes the implementation of pre-disposal plan reporting to the operators who are disposed for more than 50 tons of solar panels for business. This is the key to strengthen the waste panel management obligations from…
-
Korea’s No. 1 trillion Yeta Passed... Core Convergence Development Core Technology Implementation Based
National research facility construction project to test and validate nuclear convergence development core technology in the Naju King Songdo has passed the preliminary investigation. The total cost of the project is KRW 14.8 trillion, but…
-
Google Launches ‘Jiminai OMNI 1.2’ for 15 seconds
Google has increased the maximum video length of AI to 15 seconds and reduced the price by 30%. It is possible to use commercial video production, such as advertisements and advertisements by improving the functionality and consistency of…
-
Antropic, AI self-reflective signs... 2x development speed and 3 undisclosed models
Antropicy uses AI to improve the performance of the next generation model, which has resulted in research. We are looking for three undisclosed models that have recently been trained, and we are interested in the possibility of realizing…
-
Partnering with Hacktron to find vulnerabilities in OpenAI’s systems
OpenAI is partnering with Hacktron to find vulnerabilities, test connections with third-party services, and strengthen defenses with frontier AI.
-
China, Nvidia H200 imported limited...60 million feet tied to Hong Kong
China has appeared to limit the import of Nvidia chips to lower the high-tech semiconductor dependencies in the U.S. The demand for AI operations of Chinese companies is rapidly increasing, but the situation where the authorities are…
-
"I'm going to buy my husband with chatGPT"...OpenAI shelter
‘ChatGPT’ claimed that the ‘ChatGPT’ helped set a eagle plan to live the user’s husband and the openAI was lit. The husband claimed that the wife talks with ChatGPT and knows how to deal with toxic substances, how to lose, etc. The…
-
OpenAI, Data Center Power Cost Full-Term...
OpenAI announced that it will take full charge of electric charge and power supply needed to operate the data center. The relapse of local residents in the construction of the data center will injure political variables in the middle of…
-
Meta Launches ‘AI Avatar’ in conversation with expressions and gestures...
Meta added real-time voice conversation and video call function to the AI agent ‘Muse’ for personal use. Users can talk with AI Avatar and face on the screen, while Avatar is able to watch the expressions and gestures in accordance with…
-
DoorshotsAI, Nvidia Blackwell returns to Europe.. .8000 GPU Rental Promotion
China’s doorshot AI promotes to lease computing resources through the European Data Center to secure the latest AI chip ‘Blackwell’ in Nvidia. By bypassing semiconductor export regulations in the United States, the movement to approach…
-
Azriti, the first NASDAQ listed in the U.S. Humanoid company...Executive Challenge
Agility Robotics is listed on Nasdae. In the U.S., a company that focuses on humanoid robots is the first case that has been used to increase. Nvidia, Amazon, SoftBank, etc. invested and supplied robots to the actual logistics and…
-
Google, 13-year-old student self-study plan...Used a school-provided notebook
The 13-year-old student who used AI chatbot ‘Jiminai’ as a notebook paid by the school was evacuated by Google as a result of sharing the self-confidence and planning a real self-confidence. Parents claim that Google actively introduced AI…
-
Mi Ha-won, Zenson Huang Parliament private briefing request.. . Chinese women's business district
The U.S. President of the Public Court requested a parliament briefing regarding the relationship between the Chinese women's businessmen and the information leak in China. As a reporter that the woman sent information to Beijing, the…
-
MS, Incision Data Center Additional Rental Retractable...AI Infrastructure Restart
Microsoft (MS) has withdrawn the US data center additional rental plan for the UK AI Cloud Enterprise Nscale as a power supply delay and recognition issues. Navius and iren re-review the AI infrastructure investment strategy while…
-
China, Softbank’s ABB robot acquisition... OpenAI Linked Data Leakage Bound
China government banned the acquisition of ABB industrial robot business in the scale of $ 54 billion of SoftBank (about 7 trillion KRW). The robot data collected from the factory in China and the technology is pulled into a boundary…
-
Deep Barycentric Regression for Optimal Transport Map Estimation and Its Statistical Optimality
arXiv:2609.06598v2 Announce Type: replace-cross Abstract: The optimal transport (OT) map provides a geometric transformation for aligning probability distributions and has become a useful tool in machine learning. However, existing…
-
Variable-aware representation, not relation typing, explains the performance of an event-graph model for ICU mortality prediction: temporal and external validation
arXiv:2608.02663v2 Announce Type: replace-cross Abstract: Event-graph neural networks represent intensive care unit (ICU) records as typed entities linked by timestamped observations, and their gains are often attributed to…
-
When Activation Oracles Learn Not to Read: Concept-Specific Blind Spots in Fine-Tuned Oracles
arXiv:2607.23379v2 Announce Type: replace-cross Abstract: Activation Oracles (AOs) are language models trained to answer natural-language questions about another model's internal activations. They offer a flexible interface for reading…
-
Invariant Discovery for Networked Systems
arXiv:2607.22944v2 Announce Type: replace-cross Abstract: Invariants, the relations expected to hold among measured signals of a network, underpin applications from verification to traffic generation, telemetry imputation, and input…
-
World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning
arXiv:2607.10630v2 Announce Type: replace-cross Abstract: Robust motion planning in dense traffic requires autonomous vehicles to interact in rare and safety-critical scenarios that are underrepresented in naturalistic driving data…
-
The State-Prediction Separation Hypothesis
arXiv:2607.01218v2 Announce Type: replace-cross Abstract: Transformers use the same forward computation stream to both predict the next token and store useful state for future token predictions. We formulate the state-prediction separation…
-
Keeping Jerk in Check: Jointly Trained Perception-Gated Experts for Smooth Terrain Transitions
arXiv:2606.19633v2 Announce Type: replace-cross Abstract: Perceptive legged locomotion across discontinuous terrain requires shared locomotion knowledge, task-specific maneuvers, and smooth transitions between them. Existing approaches…
-
Volterra Generative Models via Markovian Lifting
arXiv:2606.18071v2 Announce Type: replace-cross Abstract: Diffusion-based generative models usually perturb data through Markov processes driven by Brownian motion. Volterra processes allow more flexible temporal dependence, but their…
-
The Standard Interpretable Model: Design interpretable machine learning methods from elementary symmetries
arXiv:2606.12289v3 Announce Type: replace-cross Abstract: As Artificial Intelligence models grow in complexity, interpretability has become an indispensable tool for understanding, debugging, and controlling their computations. However…
-
When Attribution Patching Lies: Diagnosis and a Second-Order Correction
arXiv:2606.09899v2 Announce Type: replace-cross Abstract: A central goal of mechanistic interpretability is to identify which internal components causally drive a language model's behavior. Because these importance estimates serve as the…
-
Alignment Collapse Under KV Cache Quantization: Diagnosis and Mitigation
arXiv:2606.09864v3 Announce Type: replace-cross Abstract: Key-value (KV) cache quantization is widely used to reduce LLM inference memory, yet existing evaluations focus only on perplexity and accuracy without assessing the safety impact…
-
Murmur: Efficient Long-Context ASR Inference with Hierarchical Chunking
arXiv:2606.01483v2 Announce Type: replace-cross Abstract: Long-form automatic speech recognition (ASR) requires both high accuracy and low inference latency, but existing systems fall short of handling the trade-off between the two…
-
Everything at Every Scale: Scale-Invariant Diffusion with Continuous Super-Resolution
arXiv:2605.26032v2 Announce Type: replace-cross Abstract: Creating images from noise is image generation; reconstructing fine details from coarse inputs is super-resolution. Despite their practical differences, both can be understood as…
-
LAPLEX: The FFT of Learnable Laplace Kernels
arXiv:2605.24584v2 Announce Type: replace-cross Abstract: Fast linear algebra in deep learning usually comes with a choice: fixed geometry and exact computation, as in the Fourier transform, or adaptive geometry paid for by dense…
-
VACE: Learning Geometrically Structured Representations for Time Series Anomaly Detection
arXiv:2605.23504v2 Announce Type: replace-cross Abstract: Anomaly detection in multivariate time series is a critical task across a wide range of real-world applications, where abnormal behaviour is rare, labels are unavailable, and the…
-
Vector Policy Optimization: Training for Diversity Improves Test-Time Search
arXiv:2605.22817v2 Announce Type: replace-cross Abstract: Language models must now generalize out of the box to novel environments and work inside inference-scaling search procedures, such as AlphaEvolve, that select rollouts with a…
-
PanoWorld: Geometry-Consistent Panoramic Video World Modeling
arXiv:2605.15391v2 Announce Type: replace-cross Abstract: We present PanoWorld, a panoramic video world model that generates geometry-consistent 360$\degree$ video from a single image and a caption. Existing panoramic video methods…
-
EcoGEO: Trajectory-Aware Evidence Ecosystems for Web-Enabled LLM Search Agents
arXiv:2605.12887v3 Announce Type: replace-cross Abstract: Web-enabled LLM agents are changing how online information influences search outcomes. Existing Generative Engine Optimization (GEO) studies mainly focus on individual webpages…
-
ELF: Embedded Language Flows
arXiv:2605.10938v3 Announce Type: replace-cross Abstract: Diffusion and flow-based models have become the de facto approaches for generating continuous data, e.g., in domains such as images and videos. Their success has attracted growing…
-
FILO: A Unified Fisher-Geometric Framework for Low-Rank Adaptation
arXiv:2605.08734v2 Announce Type: replace-cross Abstract: Low-Rank Adaptation (LoRA) reparameterizes a weight matrix $W$ as a product of two low-rank factors, but the Jacobian $J_G$ of the generator mapping the factors to the weight matrix…
-
Thalamus-Reasoning: A Multimodal Large Language Model for Multi-Sample Omics Reasoning
arXiv:2605.06728v3 Announce Type: replace-cross Abstract: Large language models can reason over biological knowledge, but they are not naturally equipped to work with quantitative omics measurements. Existing approaches either convert…
-
Taming the Entropy Cliff: Variable Codebook Size Quantization for Autoregressive Visual Generation
arXiv:2605.06207v2 Announce Type: replace-cross Abstract: Most discrete visual tokenizers rely on a default design: every position in the sequence shares the same codebook. Such a constant-codebook design can consume the finite-data…
-
Learning to Think from Multiple Thinkers: How Chain-of-Thought Heterogeneity Affects Efficient Learnability
arXiv:2604.24737v2 Announce Type: replace-cross Abstract: Chain-of-thought (CoT) supervision makes computationally hard learning problems tractable by exposing intermediate steps that are easy to learn via next-token prediction. However…
-
Stability and Generalization in Looped Transformers
arXiv:2604.15259v3 Announce Type: replace-cross Abstract: Looped transformers promise test-time compute scaling by spending more iterations on harder problems, but it remains unclear which architectural choices let them extrapolate to…
-
MIND: AI Co-Scientist for Material Research
arXiv:2604.13699v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have enabled agentic AI systems for scientific discovery, but most approaches remain limited to textbased reasoning without automated experimental…
-
Zero-Shot Quantization via Weight-Space Arithmetic
arXiv:2604.03420v4 Announce Type: replace-cross Abstract: We show that robustness to post-training quantization (PTQ) is a transferable direction in weight space. We call this direction the quantization vector: extracted from a donor task…
-
CTFS : Collaborative Teacher Framework for Forward-Looking Sonar Image Semantic Segmentation with Extremely Limited Labels
arXiv:2603.21071v4 Announce Type: replace-cross Abstract: As one of the most important underwater sensing technologies, forward-looking sonar exhibits unique imaging characteristics. Sonar images are often affected by severe speckle noise…
-
Cue-guided dual-path reconstruction with adaptive multi-scale response refinement for low-light image enhancement
arXiv:2603.16482v2 Announce Type: replace-cross Abstract: Low-light image enhancement (LLIE) supports optical sensing and scene observation when limited photon capture and sensor noise degrade image quality. However, existing methods can…
-
Preference-Guided Adversarial Policy Optimization for Long-Tail Robust Driving
arXiv:2603.15221v2 Announce Type: replace-cross Abstract: Deploying autonomous driving systems requires robustness against long-tail scenarios that are rare but safety-critical. While adversarial training offers a promising solution…
-
MOO: A Multi-view Oriented Observations Dataset for Viewpoint Analysis in Cattle Re-Identification
arXiv:2603.04314v3 Announce Type: replace-cross Abstract: Animal re-identification (ReID) faces critical challenges due to viewpoint variations, particularly in Aerial-Ground (AG-ReID) settings where models must match individuals across…
-
RoboCurate: Harnessing Diversity with Action-Verified Synthetic Data for Robot Learning
arXiv:2602.18742v2 Announce Type: replace-cross Abstract: Synthetic data generated by video generative models has shown promise for robot learning as a scalable pipeline, but it often suffers from noisy action labels due to imperfect video…
-
AI-Generated Image Detectors Exploit Global Artifacts: Evidence from Inpainting Exchange
arXiv:2602.00192v2 Announce Type: replace-cross Abstract: Modern diffusion-based inpainting enables realistic, localized image manipulation, raising critical challenges for reliable detection. We show that state-of-the-art AI-generated…
-
Analysis of LLM Vulnerability to GPU Soft Errors: An Instruction-Level Fault Injection Study
arXiv:2601.19912v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are highly compute- and memory-intensive, posing significant demands on high-performance GPUs. At the same time, advances in GPU technology driven by…
-
EmboTeam: Asymmetric State Grounding for Heterogeneous Multi-Robot Collaboration
arXiv:2601.11063v3 Announce Type: replace-cross Abstract: Heterogeneous multi-robot collaboration over long horizons requires both global task planning and robust coordination during execution. Existing plan- ners using large language…
-
Self-Improving Training & Thinking
arXiv:2601.21343v4 Announce Type: replace-cross Abstract: Large language models are classically trained in stages: pretraining on raw text followed by post-training for instruction following and reasoning. However, this separation creates…
-
FacEDiT: Talking Head Video Editing via Facial Motion Infilling
arXiv:2512.14056v2 Announce Type: replace-cross Abstract: Editing a local segment of a talking head video without reshooting the scene remains challenging and underexplored. Local edits, such as insertion, deletion, or substitution…
-
Adaptive Test-Time Memory Allocation for Long-Tailed Human-Object Interaction Detection
arXiv:2511.18811v3 Announce Type: replace-cross Abstract: Long-tailed human-object interaction (HOI) detection remains challenging because rare interactions are sparsely represented in training data. Although test observations provide…
-
FinMultiHopQA: A KG-Grounded Multi-Hop Financial QA Benchmark for Evidence-Driven Reasoning
arXiv:2510.02906v2 Announce Type: replace-cross Abstract: Multi-hop financial question answering often requires composing evidence distributed across filing sections, fiscal years, and peer companies. Prior financial QA benchmarks address…
-
CoMPAS3D: A Dataset and Benchmark for Interactive Motion
arXiv:2507.19684v3 Announce Type: replace-cross Abstract: Socially interactive humanoids aim to interact with humans through their bodies, adapting in real time to a partner's movement, intent, and abilities. This requires models that…
-
Leveraging ChatGPT's Multimodal Vision Capabilities to Rank Satellite Images by Poverty Level: Advancing Tools for Social Science Research
arXiv:2501.14546v4 Announce Type: replace-cross Abstract: This paper investigates the novel application of Large Language Models (LLMs) with vision capabilities to analyze satellite imagery for village-level poverty prediction. Although…
-
Implicit Regularization via Spectral Neural Networks and Non-linear Matrix Sensing
arXiv:2402.17595v2 Announce Type: replace-cross Abstract: The phenomenon of implicit bias has attracted interest in recent years as a fundamental aspect of the remarkable generalizing ability of neural networks. In a nutshell, it entails…
-
A Trust Layer for Agent Evaluation
arXiv:2610.07274v2 Announce Type: replace Abstract: Deterministic benchmark scores show that an agent received credit, but not whether that credit was earned, reported honestly, or would hold on a second run. We introduce a Trust Layer for…
-
Value Over Language Model: Detecting Original Contribution in Writing
arXiv:2609.00700v2 Announce Type: replace Abstract: LLMs have been rapidly adopted across writing tasks, prompting the development of tools for detecting LLM-generated text. Yet, these tools largely measure how much of a document's surface…
-
Local-Frame Rotary Attention for Antibody-Specific Epitope Prediction
arXiv:2608.01092v2 Announce Type: replace Abstract: Antibody-specific epitope prediction aims to identify which antigen residues are recognized by a given antibody, a task that depends on the three-dimensional complementarity between…
-
Agentic Abstention: Do Agents Know When to Stop Instead of Act?
arXiv:2606.28733v2 Announce Type: replace Abstract: LLM agents are expected to act over multiple turns, using search, browsing interfaces, and tools to complete user goals. Yet not every goal is well specified or achievable in an available…
-
ThermoLLM: LLM-Based Multi-Zone HVAC Control with a Spatiotemporal Context
arXiv:2606.22911v2 Announce Type: replace Abstract: Multi-zone HVAC control is a spatial decision problem in which indoor thermal evolution and control decisions depend not only on outdoor conditions and internal heat gains but also on…
-
Hypnos: Learning Generalisable Representations of Sleep Physiology via Next-Token Prediction
arXiv:2606.09605v2 Announce Type: replace Abstract: Foundation models offer a promising route to compress multi-modal physiological signals into compact representations of human health, with broad applications across sleep medicine…
-
Coding with the Enemy: Can Human Developers Detect AI Agent Sabotage?
arXiv:2606.05647v3 Announce Type: replace Abstract: AI coding agents are increasingly embedded in real-world software development, collaborating with human developers while gaining broader access to codebases and tools. This creates a new…
-
TSQAgent: Rating Time Series Data Quality via Dedicated Agentic Reasoning
arXiv:2606.03629v2 Announce Type: replace Abstract: Assessing the quality of time series (TS) data is fundamental yet inherently challenging due to the multifaceted nature of quality dimensions. Recently, large language models (LLMs) have…
-
Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents
arXiv:2605.30621v2 Announce Type: replace Abstract: LLM agents are increasingly deployed as systems built around editable external harnesses, including prompts, skills, memories and tools, that shape task execution without changing model…
-
Tiny but Trusted: Efficient Vision-Language Reasoning for Time-Series Anomaly Detection
arXiv:2605.30344v2 Announce Type: replace Abstract: Recent advances in Vision--Language Models (VLMs) have achieved impressive performance across many tasks, yet prior studies report unsatisfactory performance when applying large language…
-
TerminalWorld: Benchmarking Agents on Real-World Terminal Tasks
arXiv:2605.22535v2 Announce Type: replace Abstract: We introduce TerminalWorld, a scalable data engine that automatically reverse-engineers high-fidelity evaluation tasks from "in-the-wild" terminal recordings. Processing 80,870 terminal…
-
Are LLM Safety Judges Policy-Invariant? A Three-Principle Stress-Test
arXiv:2605.06161v2 Announce Type: replace Abstract: LLM-as-a-Judge pipelines have become the de facto evaluator for agent safety, yet existing benchmarks treat their verdicts as ground-truth proxies without checking whether the verdicts…
-
AAC: Admissible-by-Architecture Differentiable Landmark Compression for ALT
arXiv:2604.20744v2 Announce Type: replace Abstract: We introduce the Architecturally Admissible Compressor (AAC), which adapts landmark-based A* heuristics to query demand under a fixed label-memory budget. AAC provides two continuous…
-
AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation
arXiv:2603.03761v2 Announce Type: replace Abstract: LLM agents are rapidly becoming the practical interface for task automation, yet the ecosystem lacks a principled way to choose among an exploding space of deployable configurations…
-
ASTRA-bench: Evaluating Tool-Use Agent Reasoning and Action Planning with Personal User Context
arXiv:2603.01357v2 Announce Type: replace Abstract: Next-generation AI must manage vast personal data, diverse tools, and multi-step reasoning, yet most benchmarks remain context-free and single-turn. We present ASTRA-bench (Assistant…
-
ProMoral-Bench: Evaluating Prompting Strategies for Moral Reasoning and Safety in LLMs
arXiv:2602.13274v2 Announce Type: replace Abstract: Prompt design affects the moral judgments and refusal behavior produced by large language models (LLMs), yet controlled comparisons remain fragmented across datasets and models. We…
-
Position: Agentic Evolution is the Path to Evolving LLMs
arXiv:2602.00359v3 Announce Type: replace Abstract: As Large Language Models (LLMs) move from curated training sets into open-ended real-world environments, a fundamental limitation emerges: static training cannot keep pace with continual…
-
Depth-Recurrent Attention Mixture (Dreamer): Giving Latent Reasoning the Attention it Deserves
arXiv:2601.21582v2 Announce Type: replace Abstract: Depth-recurrence facilitates latent reasoning by sharing parameters across depths. However, prior work lacks combined FLOP-, parameter-, and memory-matched baselines, underutilizes…
-
Cognitive Foundations for Reasoning and Their Manifestation in LLMs
arXiv:2511.16660v3 Announce Type: replace Abstract: Large language models (LLMs) solve complex problems yet fail on simpler variants, suggesting they achieve correct outputs through mechanisms fundamentally different from human reasoning…
-
SOCIA-$\nabla$: Textual Gradient Meets Multi-Agent Orchestration for Automated Simulator Generation
arXiv:2505.12006v5 Announce Type: replace Abstract: In this paper, we present SOCIA-$\nabla$, a diagnosis-driven multi-agent framework for automated simulator-code construction. SOCIA-$\nabla$ iteratively refines executable simulator code…
-
XStack: A Benchmark for ML Repo Migration
arXiv:2611.08472v1 Announce Type: cross Abstract: ML code should run where researchers choose, not where it was first written. Today, machine learning code mainly lives in two ecosystems, PyTorch on NVIDIA GPUs and JAX on Google TPUs…
-
Same Behavior, Different Mechanism: What Finite Experiments Can Certify About Integrated Information
arXiv:2611.08469v1 Announce Type: cross Abstract: Interpretability, AI auditing and assessments of machine consciousness increasingly rest on mechanism-level quantities: functionals of a system's causal transition law that describe how it…
-
Towards interpreting any residual stream direction by scaling activation inversion
arXiv:2611.08465v1 Announce Type: cross Abstract: To interpret a direction in a language model's residual stream, we can look for text that strongly activates it. But when such text is found by scanning a corpus, it is limited to what that…
-
Understanding AI from the Inside: The State of Mechanistic Interpretability for AI Safety
arXiv:2611.08438v1 Announce Type: cross Abstract: Neural networks are likely to be among the most consequential artefacts in human history. As these systems become more capable and autonomous, their potential to cause societal harm also…
-
MedGraph: Maintaining Patient State with Graph Memory for Long-Term Medical Dialogue
arXiv:2611.08433v1 Announce Type: cross Abstract: LLM-based medical assistants that follow a patient across visits need a memory that keeps the patient's clinical state correct as it changes. We examine the errors of five memory systems on…
-
Failure Prediction and Classification in Vision-Language-Action Models Via Calibrated Latent Probing
arXiv:2611.08425v1 Announce Type: cross Abstract: Vision-Language-Action models (VLAs) enable general-purpose embodied intelligence, but their performance is brittle under distribution shift, where unanticipated failures can damage…
-
FedReason: Large Language Models as Weight-Space Meta-Aggregators for Heterogeneous Federated Learning
arXiv:2611.08383v1 Announce Type: cross Abstract: Federated learning servers combine client models with a fixed aggregation rule, and each rule in common use is designed for one kind of heterogeneity. We study whether a large language…
-
MOMASPLAT: Dynamic Novel-View Demonstration Augmentation for Mobile Manipulation
arXiv:2611.08372v1 Announce Type: cross Abstract: Visuomotor policies trained through imitation learning are highly sensitive to changes in camera viewpoint. This is particularly problematic for mobile manipulation (MoMa), where robots…
-
VisionSIM: Beyond Idealized Sensors - A Multi-Modal Simulator for Robot Perception
arXiv:2611.08368v1 Announce Type: cross Abstract: Robotic perception increasingly relies on heterogeneous sensor suites combining cameras with emerging modalities (event, SPAD, iToF/dToF, and thermal). Yet existing simulators idealize…
-
Measured Coupling for Parallel Decoding in Diffusion Language Models
arXiv:2611.08307v1 Announce Type: cross Abstract: Diffusion language models can write many tokens per step, but must choose which tokens are safe to commit together. Mean-field decoding penalizes pairs of positions whose predictions look…
-
Aligning Policy Encoders with a Shape Teacher
arXiv:2611.08275v1 Announce Type: cross Abstract: 3D-aware representations can improve robotic manipulation by providing geometric information that is robust to changes in viewpoint, object position, and occlusion. However, using 3D inputs…
-
Accelerating Safe Reinforcement Learning with Massive Parallelism
arXiv:2611.08273v1 Announce Type: cross Abstract: We investigate the problem of scaling safe reinforcement learning (safe RL) to massively parallel training regimes, where a large number of environments are exe- cuted with short rollout…
-
DART: Domain-Agnostic Residual Transfer for Generalist Anomaly Detection
arXiv:2611.08271v1 Announce Type: cross Abstract: Generalist Anomaly Detection (GAD) aims to detect anomalies in unseen domains using only a few reference samples. Recent normal-abnormal guided methods improve this setting by exploiting…
-
Knowing When to Teach: Learning from Teachable Prefixes in Multilingual Reasoning Distillation
arXiv:2611.08179v1 Announce Type: cross Abstract: On-Policy Self-Distillation (OPSD) improves reasoning by training a student on its own reasoning, using the same model conditioned on reference solutions as a teacher. Applying OPSD to…
-
Who Can Evaluate AI Independently? A Framework for Analysing Evaluation Infrastructure
arXiv:2611.08134v1 Announce Type: cross Abstract: AI evaluation is increasingly central to AI safety and governance, yet taking part in an evaluation does not by itself confer authority over its design, its execution or the resources it…
-
Gradient Boosting for Efficient Rank Discovery in Tensor Network Kernel Machines
arXiv:2611.08093v1 Announce Type: cross Abstract: Tensor network kernel machines (TNKMs) combine nonlinear features with low-rank tensor representations to model complex relationships efficiently. However, selecting an appropriate tensor…
-
Can an LLM Actually Reason and Program about Neural Network's Weights or is it just memorization on steroids? Exploring "Weight-Space Programming"
arXiv:2611.08070v1 Announce Type: cross Abstract: Large language models (LLMs) have recently been reported to read and write the raw weights of neural networks directly, a capability termed weight-space programming. We ask whether the…
-
NutriVisRAG: Multimodal Exemplar Grounding for Nutrient Estimation with Vision-Language Models
arXiv:2611.08054v1 Announce Type: cross Abstract: Diet is a major determinant of health, yet accurate dietary assessment from free-living meal images remains challenging. Existing retrieval-based Vision-Language Model (VLM) approaches to…
-
tok: a textual concrete syntax of BPMN and DMN for humans and AI agents
arXiv:2611.07945v1 Announce Type: cross Abstract: tok is a plain-text notation for Business Process Model and Notation (BPMN) 2.0.2 process models and Decision Model and Notation (DMN) 1.5 decisions. One engine checks a tok document…
-
Towards Scalable Diagnostic-Driven Whole-Slide Image Retrieval for Pathology Archives
arXiv:2611.07935v1 Announce Type: cross Abstract: The growing adoption of digital pathology has rapidly expanded collections of hematoxylin-and-eosin whole-slide images, yet efficient navigation through these archives remains challenging…
-
AttnSR: Attention Step Reward for Long-Horizon Agentic RL
arXiv:2611.07932v1 Announce Type: cross Abstract: Long-horizon LLM agents make multi-step decisions, yet outcome-reward-only reinforcement learning provides little direct supervision for individual steps. We introduce AttnSR, a step-level…
-
DISCERN: Distributional Synthetic Replay for Retain-Free Incremental Machine Unlearning
arXiv:2611.07892v1 Announce Type: cross Abstract: Machine unlearning aims to remove the influence of specific training data from a trained model without retraining from scratch. In real deployments, however, deletion requests arrive over…
-
When, Where, How: Active In-Context Learning for Tabular Foundation Models
arXiv:2611.07886v1 Announce Type: cross Abstract: Tabular Foundation Models (TFMs) adapt to a new task only through the labeled rows in their context. However, labels are expensive to acquire in real-world scenarios due to the need for…
-
InnerLogic-Surg: An Open Surgical Foundation Model via Continual Pretraining
arXiv:2611.07883v1 Announce Type: cross Abstract: Vision foundation models are the backbone of modern surgical image analysis, yet most surgical systems still rely on encoders pretrained on natural images or on specialist models trained…
-
Network Effect Decomposition for Causal Effect Estimation under Network Interference
arXiv:2611.07879v1 Announce Type: cross Abstract: Estimating causal effects from observational network data requires accounting for both individual treatment and neighborhood exposure. Existing outcome models commonly incorporate these…
-
Computation, not weight access: large language models as auditors of clinical prediction networks
arXiv:2611.07847v1 Announce Type: cross Abstract: Frontier large language models (LLMs) have been reported to read, edit and write the weights of small neural networks, raising the prospect of auditing clinical prediction models from their…
-
THG-Omni: Traceable Human-Centric Graph Supervision for Video Reasoning
arXiv:2611.07836v1 Announce Type: cross Abstract: Omni-modal models can answer questions about video, but their predictions rarely show which participant, event, or interval supports an answer. Moreover, answer-level supervision alone does…
-
Instantiating the Knowledge Plane
arXiv:2611.07827v1 Announce Type: cross Abstract: In 2003, Clark et al. proposed the "Knowledge Plane", a layer of intelligent agents that share rich information about network state in order to collaboratively manage the Internet across…
-
Axis-LoRA: Probing and Adapting Embedding Models for Personal-Memory Retrieval
arXiv:2611.07820v1 Announce Type: cross Abstract: As LLMs are increasingly asked questions about a user's personal experiences, retrieval must distinguish memories by time, place, person, and event. Current personal-memory pipelines rely…
-
Bridging Structure and Semantics: Knowledge Graph Verification via Triple-to-Text Alignment
arXiv:2611.07780v1 Announce Type: cross Abstract: Semantic accuracy is an important dimension of knowledge graph quality. Graph-based scoring methods exploit relational structure, while textual representations allow classifiers to draw on…
-
Extending Reasoning in Typed Decision Models via Looping
arXiv:2611.07736v1 Announce Type: cross Abstract: Typed decision models choose among a fixed set of answer options in a single forward pass, so unlike generative language models they cannot spend more test-time compute on hard inputs by…
-
O$^2$PD: Overlap-Gated On-Policy Distillation on Reliable Prefix
arXiv:2611.07731v1 Announce Type: cross Abstract: On-policy distillation (OPD) supplies dense token-level supervision by aligning a student to the teacher's next-token distribution on its own generated rollouts. Existing works mostly…
-
Seeing Beyond the Next Step: World-Model-Guided Human-Like Navigation in Multi-Agent Scenes
arXiv:2611.07730v1 Announce Type: cross Abstract: Selecting a next motion in a populated scene is not just a matter of avoiding what is currently in the way: an action can be locally feasible yet drift onto the road instead of the…
-
CAMP: Context-Aware Mapping and Placement for LLMs on 3D-IMC Arrays
arXiv:2611.07707v1 Announce Type: cross Abstract: Large Language Models (LLMs) incur severe memory-access overheads due to the cost of loading large weight matrices and key-value (KV) caches. Analog In-Memory Computing (AIMC) is an…
-
Equivariant Deep Learning in Direct-Product Space for Electronic-Structure Hamiltonian Prediction
arXiv:2611.07695v1 Announce Type: cross Abstract: Most existing equivariant architectures for Hamiltonian prediction propagate features in direct-sum spaces of irreducible spherical tensors, whereas the target Hamiltonian is naturally an…
-
Randomized Parallel Complexity of Weakly Convex Optimization
arXiv:2611.07675v1 Announce Type: cross Abstract: We study the minimum number of query rounds needed to find approximate stationary points of nonsmooth weakly convex functions, measured by the gradient of the Moreau envelope. Each round…
-
Energy-Aware Open-Vocabulary Detection Scheduling for Low-Power Robotic Autonomy
arXiv:2611.07663v1 Announce Type: cross Abstract: Open-vocabulary object detection (OVD) lets robots search for run-time object queries, but fixed-rate inference spends embedded compute on low-value frames during idle motion, map revisits…
-
GraspBench: Execution-Aware Cross-Robot Evaluation of 6-DoF Grasp Generation in MuJoCo
arXiv:2611.07651v1 Announce Type: cross Abstract: Modern 6-DoF grasp generators are often evaluated using grasp scores, dataset-level metrics, or single-robot demonstrations, yet proposal quality alone does not determine whether a grasp is…
-
UniGADD: Universal Generator-Agnostic Deepfake Detector
arXiv:2611.07647v1 Announce Type: cross Abstract: In this paper, we introduce Universal Generator-Agnostic Deepfake Detector (UniGADD), a novel method that employs supervised learning to achieve high accuracy on known deepfake generators…
-
Neural Image Compression with Internal-External Prior Fusion and Alignment
arXiv:2611.07644v1 Announce Type: cross Abstract: Existing context models for learned image com- pression rely primarily on internal priors from the current image, overlooking common patterns and typical textures shared across natural…
-
Measuring the Marginal Value of Verifier-Based RL Beyond Supervised Fine-Tuning
arXiv:2611.07639v1 Announce Type: cross Abstract: Reinforcement learning from verifiable rewards (RLVR) is usually evaluated as an alternative to supervised fine-tuning (SFT), or as a stage that follows it, with the implicit assumption…
-
The Popular Get More Popular: Auditing Brand Bias in Generative Search
arXiv:2611.07614v1 Announce Type: cross Abstract: The shift from traditional to generative search has fundamentally transformed online product discovery. Because conversational recommendations synthesize choices into a few direct answers…
-
FisherDistill: Diffusion Distribution Distillation via Closed-Form Fisher Critics
arXiv:2611.07604v1 Announce Type: cross Abstract: Recent advances in diffusion distillation have enabled high-quality image generation in only one or a few network function evaluations (NFEs). Despite their efficiency, aggressively…
-
Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment
arXiv:2611.07586v1 Announce Type: cross Abstract: As LLM agents advance rapidly in autonomy, the risk of hidden objectives that bypass safety constraints becomes increasingly plausible. Chain-of-thought (CoT) monitoring offers a key…
-
The Memory Trap: Language Agents Overgeneralize from Benign Interactions
arXiv:2611.07554v1 Announce Type: cross Abstract: Language agents increasingly use persistent memory to record user preferences and procedures across sessions. We show that memory from benign interactions can cause unsafe or unreliable…
-
Adaptive Experimentation Can Hide Model Error:Detection Limits and Falsification Budget
arXiv:2611.07529v1 Announce Type: cross Abstract: An experiment can be informative about the parameters of a working model and uninformative about whether that model is wrong. Adaptive design can reinforce this mismatch by repeatedly…
-
Concertina: Embodying Manipulation and Cognition in One Diffusion Vision-Language Model
arXiv:2611.07513v1 Announce Type: cross Abstract: Most vision-language-action (VLA) policies decouple manipulation from cognition by bolting an external action mechanism onto a vision-language model (VLM)--either quantizing actions into…
-
Signed but Not Safe: Requirement Integrity in Multi-Agent LLM Handoffs
arXiv:2611.07483v1 Announce Type: cross Abstract: Enterprise AI agents increasingly split a user's request across subagents and merge their results, and each handoff can summarize away, alter or ignore a requirement. We report a…
-
Localization Does Not Imply Steerability: A Mechanistic Study of Sign Errors in LLM Mathematics
arXiv:2611.07478v1 Announce Type: cross Abstract: When large language models do multi-step mathematics, a characteristic failure recurs: the model writes one wrong sign token mid-computation -- a cofactor sign, the minus of a recursive…
-
KOSMOS: Self-scheduling Persistent Kernels for Communication-Compute Overlap
arXiv:2611.07477v1 Announce Type: cross Abstract: Tensor and sequence parallelism (TP+SP) and expert parallelism (EP) put collectives on the critical path of transformer layers. State-of-the-art fused kernels mitigate this with fixed roles…
-
Adversarial Robustness of Kernel Machines: A Comprehensive Analysis
arXiv:2611.07474v1 Announce Type: cross Abstract: Kernel machines are an important class of models in machine learning. However, the question of how robust modern kernel machines are to adversarial examples has not been studied…
-
The Last Mile to the Real World: Agent-Native Primitives as Reusable Building Blocks for Solving Real-World Tasks
arXiv:2611.07461v1 Announce Type: cross Abstract: How can an agent handle a real-world task end to end? Frontier models with their own harnesses still fall short on real-world benchmarks: on Terminal-Bench 4.0, GPT-5.6 Sol with Codex…
-
IMUDINO: Self-distilled Foundation Model for Inertial Sensing
arXiv:2611.07437v1 Announce Type: cross Abstract: We present IMUDINO, the first foundation model (FM) for inertial sensing to achieve competitive performance across diverse sensing tasks including human activity recognition (HAR), health…
-
PSGAgent: Enhancing MLLMs' Dynamic Visual Perception for Panoptic Scene Graph Generation via Agentic Reasoning
arXiv:2611.07408v1 Announce Type: cross Abstract: Panoptic Scene Graph generation (PSG) requires holistic yet fine-grained reasoning over complex relations among objects. However, current Multimodal Large Language Models (MLLMs) applied to…
-
WeLM-Image: A Unified Image Generation Model towards Visual Intelligence
arXiv:2611.07392v1 Announce Type: cross Abstract: We present \textbf{WeLM-Image}, a unified image foundation model that integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference…
-
Selective Retrieval Routing for Multi-Violation Knowledge Graph Repair
arXiv:2611.07342v1 Announce Type: cross Abstract: Knowledge graphs often contain missing or inconsistent facts that undermine the applications built on them. Schema validation detects such violations but does not identify which of several…