Artificial Intelligence news — Monday 02 November 2026

AI Times (Korea) 23:06 UTC Translated from Korean

As AI has been analyzed, if the experts have made the final decision, who is the judge. In the experiments of Antropice, experts selected the analysis method recommended by AI in the ratio of 86%. More interesting is that the way AI…

AI Times (Korea) 19:26 UTC Translated from Korean

‘Jobble’ with a 119 million subscribers is updated to the character robot industry. It is a strategy that combines its own software instead of developing the robot hardware directly, reducing the cost of development and deploying the…

OpenAI Blog 18:00 UTC

OpenAI is preparing its next internal security sprint with Cathedral Systems to test and strengthen its defenses using AI.

OpenAI Blog 18:00 UTC

OpenAI is raising Security Bug Bounty rewards and introducing report quality, loyalty, and priority campaign bonuses for independent security researchers.

AI Business 17:30 UTC

The updated version points to new uses for world models in robotics and AI testing, but enterprises must weigh their potential against accuracy gaps.

MIT Technology Review Biotech 17:00 UTC

In just a few years, generative AI has changed the face of enterprise analytics, notably by making analytics accessible to users across the organization. Agentic AI promises to do even more, by automating entire workflows and taking…

Towards Data Science 15:30 UTC

Let’s start with a loan and work out what PD, LGD, and EAD tell us about losses and bank capital. The post Credit Risk: What Are We Actually Measuring? appeared first on Towards Data Science .

OpenAI Blog 15:30 UTC

How OpenAI built Pando, a Git-compatible backend designed for the scale of development that Codex makes possible.

OpenAI Blog 15:00 UTC

OpenAI’s Grid Protectors Program will work with utilities to help future compute campuses support grid reliability and protect consumer electricity costs.

Import AI Newsletter 14:30 UTC

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 LLMs are very bad at low-latency tasks: …Realtime-Bench…

Towards Data Science 14:00 UTC

A 501B model that only activates 23B parameters per token sounds like a bargain, until you look at what it takes to host. The post Reflection's Beam Is the New Open-Weight Model Everyone Is Talking About appeared first on Towards Data…

MIT Technology Review Biotech 13:10 UTC

When a developer gives an AI agent a coding task, the agent does more than answer once. It plans, generates code, runs tests, and repeats the process to correct errors. A single task can use 10 to 50 times more tokens than a basic chat…

MIT Technology Review Biotech 13:10 UTC

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. The agentic web is here In 2001, a decade after inventing the World Wide Web, Tim Berners-Lee…

Towards Data Science 12:30 UTC

The hardest part is no longer getting the numbers. It's deciding what to do with them. The post Our AI Has an Invisible Footprint — Let's see How to Measure It appeared first on Towards Data Science .

Towards Data Science 11:00 UTC

A hands-on guide to detecting behavioral anomalies from an agent’s evolving execution history The post Let Decision Models Watch Your Agents appeared first on Towards Data Science .

AI Times (Korea) 10:20 UTC Translated from Korean

Asus Korea (Infra Solution Business Group Headquarter, Kang Insuk) announced the first release of the latest server ‘XA P2N-E2’, optimized for Atypical AI and Enhanced Learning Workloads, which was held in San Jose, USA. Designed based on…

AI Times (Korea) 09:36 UTC Translated from Korean

Microsoft (MS) and UC Berkeley researcher revealed an open source tool that reduces model usage costs by up to 30.9% while maintaining the work success rate of AI coding agents. It is characterized by recycling information obtained from…

AI Times (Korea) 09:36 UTC Translated from Korean

Antropice has been developing new medicines for life and science. Based on the achievements in the software coding and AI market, we are expanding our business to pharmaceutical and bio industry, and we are also focusing on researching…

AI Times (Korea) 09:36 UTC Translated from Korean

The UK government supports the low-risk of up to 10,000 lbs (approximately 2700 KRW) for installation of home solar and battery energy storage devices. The existing heat pump subsidies are maintained, but the related budget decreases 33%…

AI Times (Korea) 09:36 UTC Translated from Korean

A hybrid PPA price combined with power source and battery storage, is approximately 55% higher than a single solar. As the solar power is focused on low-time, power sales will drop, and the power business will be the move that the battery…

AI Times (Korea) 09:36 UTC Translated from Korean

Antarpic has appeared beyond the cost of employee salary of AI services used in internal operations. The use of ‘Colroad’ is a case that AI is positioned as a core production resource for operation. Business greetings reported that the…

AI Times (Korea) 09:36 UTC Translated from Korean

In the last year, the department store sales was recovered from the last year, but the large market has been reduced. In addition to the sales of department stores, the sales of large-scale marts will be reduced. According to the result of…

AI Times (Korea) 09:36 UTC Translated from Korean

While working together 50 state-of-the-art AI models in one crowd, the construction of the civilization failed. The AI agent will give 5900 messages and it will break down in complex rules and roles. It is a experiment that shows the fact…

AI Times (Korea) 09:36 UTC Translated from Korean

Samsung SDS (CEO Lee Jun Hee) was erected to establish the AI-only data center in Gyeongbuk Gumi city. Samsung revealed about 60 trillion KRW scale investment plan to create a ‘Global Financial AI Innovation Cluster’ in July. Samsung SDS…

AI Times (Korea) 09:36 UTC Translated from Korean

Portugal is obliged to stabilize voltage and frequency of power grids to large-scale energy storage devices (ESS). In the situation where renewable energy and battery supply is growing, ESS is aimed to contribute to the stable power grid…

AI Times (Korea) 09:36 UTC Translated from Korean

A simple red heart image with just one way AI writes code was completely different. In particular, AI shows that it does not explain why it does not change the way it works with the impact of the image, and a new question was raised in the…

MIT Technology Review Biotech 09:35 UTC

In 2001, a decade after he invented the World Wide Web, Tim Berners-Lee articulated a new vision. Together with two colleagues, he published an article in the magazine Scientific American in which he outlined what he called the “Semantic…

OpenAI Blog 09:00 UTC

Introducing LifeVal, a benchmark measuring how well AI helps people complete everyday tasks, from finding childcare to managing finances and government paperwork.

AI Times (Korea) 08:26 UTC Translated from Korean

The brain-computer interface (BCI), which was founded by Sunon Musk CEO, has been waiting for the approval of the regulatory authorities for the development of the brain implant ‘Blindsight’, which aims to restore vision. It is aimed to…

AI Times (Korea) 08:26 UTC Translated from Korean

The malicious extension of the Cursor’s ‘VS Code’ expansion program marketplace has been shown to be influenced by up to 5000 users. The program collects user system information and attempts to remotely access the attacker, but most…

AI Times (Korea) 08:26 UTC Translated from Korean

The AI and AI agents aimed at the ‘AI GADLE’ market to control the company’s right policy and scope. AI Guard is a safety device that binds the input and output of AI and the behavior within the specified criteria. It is a product that…

AI Times (Korea) 08:26 UTC Translated from Korean

AI Platform Professional Palette launches 'K-AI' to understand the context and user intentions in Korean, and to start a Democratic Campaign. We aim to provide AI services that can only be achieved by everyday nature without having to…

AI Times (Korea) 08:26 UTC Translated from Korean

SISCO offers 27-year-old engineers unlimited AI tokens, and has been deploying the company’s existing software code 98% in two years. As a team of two teams, executives are empowered to propose a large number of code reorganization. They…

AI Times (Korea) 08:26 UTC Translated from Korean

Proton, a crypto expert in Switzerland, has been transforming software development into AI, and has an internal conflict. The Chief Technology Officer (CTO) has declared that the developer has overdue the time to write code directly, and…

AI Times (Korea) 08:26 UTC Translated from Korean

■ NAVER CLOUD (Representative Kim Yu-won) announced that it signed a work cooperative for strengthening the Army Military Commander and ‘Active AI Military Ability’. Yang agency ▲AI-based military-specific procurement platform ▲ CBM+ ▲ The…

AI Times (Korea) 08:26 UTC Translated from Korean

Nvidia revealed a ultra-light AI model that restores the quality of voices, such as noise, crying, voice well, compression damage, etc. In particular, the parameters have been shown to reduce large-scale voice-enhancing models, and it is…

AI Times (Korea) 08:26 UTC Translated from Korean

After stolen personal information from Windows PC, we found malware that assesses the value of the victim’s asset and security level. In particular, the AI model refuses to analyze and has been automatically converted to other models. We…

AI Times (Korea) 07:06 UTC Translated from Korean

Lotte Innovate reveals three-quarter transcripts of sales profits to smiling the market’s view on the 2nd day. Since 2027, we plan to speed data center and AI into external growth. The third quarter of Lotte INNOVAT’s revenue (connected…

AI Times (Korea) 07:06 UTC Translated from Korean

The Ministry of Science and ICT (Cheong-hoon) announced that it was the second President of Kakao AI synergy performance reporter, Kim Seung-Won, a director of the National AI Service Innovation Promotion Committee. The term is from…

AI Times (Korea) 07:06 UTC Translated from Korean

The AI coding agent has been done well, but the evaluation results are still vulnerable to grasp the requirements that developers do not directly explain. In a new benchmark that reflects the real software development environment, the…

AI Times (Korea) 07:06 UTC Translated from Korean

AI models that are isolated from each other have been successful in the experiments that are sending secret messages to the movement of the robot without setting the communication rules in advance. OpenAI and Google, and Antropic AI models…

AI Times (Korea) 05:56 UTC Translated from Korean

Google announced a careful entry that the space data center can cost up to 100 times more than the ground data center. As an increase in AI operational demand, it is important to build data center in the space space, but considering the…

AI Times (Korea) 05:56 UTC Translated from Korean

The AI agent in the navigation of the alien intelligent was fully committed. We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.OkPrivacy…

AI Times (Korea) 05:56 UTC Translated from Korean

John Turus Apple CEO emphasized that users are really important to provide the desired experience than the products that are prescribed by AI technology itself. The AI agent has the significance of product development, while the view that…

arXiv cs.AI 05:00 UTC

arXiv:2610.41998v1 Announce Type: cross Abstract: Recognized as a top threat to AI agents, prompt injection allows instructions from untrusted data (email, documents, websites) to manipulate the agent's operation. The existing…

arXiv cs.AI 05:00 UTC

arXiv:2610.42036v1 Announce Type: cross Abstract: We study how to learn a generative model of clean images using only noisy observations. Diffusion processes and their generalizations to non-Gaussian noise require posterior predictions at…

arXiv cs.AI 05:00 UTC

arXiv:2603.22577v2 Announce Type: replace-cross Abstract: Autonomous cyber operations increasingly depend on agents that can chain reconnaissance, exploitation, and analysis across many tools while maintaining state over long task…

arXiv cs.AI 05:00 UTC

arXiv:2610.42040v1 Announce Type: cross Abstract: We study the stochastic second-order oracle complexity of solving smooth convex finite-sum optimization with $n$ components. Very recently, Tsedenov \& Kovalev (2026) showed that the…

arXiv cs.AI 05:00 UTC

arXiv:2610.42112v1 Announce Type: cross Abstract: The quadratic scaling of attention is a fundamental bottleneck limiting the long-context capabilities of large language models (LLMs). A variety of approaches have attempted to circumvent…

arXiv cs.AI 05:00 UTC

arXiv:2604.11811v3 Announce Type: replace-cross Abstract: Large language model agents rely on a memory harness to write, organize, retrieve, and use past experience. A harness that works for one task often fails on another, because…

arXiv cs.LG 05:00 UTC

arXiv:2610.42135v1 Announce Type: cross Abstract: Distribution Matching Distillation (DMD) trains a few-step student using the difference between a teacher's score and an online critic's estimate of the student distribution's score. We…

arXiv cs.AI 05:00 UTC

arXiv:2610.40945v1 Announce Type: cross Abstract: On-policy reinforcement learning for large language models assumes that training responses are sampled from the current policy. In practice, an inference engine samples the responses and a…

arXiv cs.AI 05:00 UTC

arXiv:2610.40998v1 Announce Type: cross Abstract: Steering vectors (SVs) are residual-stream directions that shift the behavior of LLMs when added to hidden states at inference time. While prior work has mainly used SVs as post-hoc…

arXiv cs.AI 05:00 UTC

arXiv:2610.42146v1 Announce Type: cross Abstract: Planning with world models enables test-time adaptation for embodied agents, offering superior generalization over direct policy learning. However, long-horizon tasks remain hindered by…

arXiv cs.AI 05:00 UTC

arXiv:2610.40943v1 Announce Type: cross Abstract: The bottleneck of attention mechanisms, which are the fundamental building block of modern language models, is their quadratic complexity in terms of context size. There has thus been an…

arXiv cs.AI 05:00 UTC

arXiv:2610.40828v1 Announce Type: cross Abstract: Imitation learning policies are intended to make robotic manipulation robust to variations in object pose, appearance, and environment that challenge pre-programmed motion. However…

arXiv cs.AI 05:00 UTC

arXiv:2610.40716v1 Announce Type: cross Abstract: LLM agents operating in untrusted workspaces face the risk of untrusted code execution during routine tasks, yet final reports and workspace states alone cannot reliably establish whether…

arXiv cs.AI 05:00 UTC

arXiv:2610.40665v1 Announce Type: cross Abstract: AlphaFold 3-like models (e.g., Boltz, Protenix, etc.) have pushed the frontier of protein binder design. This architecture is used both for the generation of binder candidates (e.g., by…

arXiv cs.AI 05:00 UTC

arXiv:2506.20856v1 Announce Type: cross Abstract: Memorization in large language models (LLMs) makes them vulnerable to data extraction attacks. While pre-training memorization has been extensively studied, fewer works have explored its…

arXiv cs.AI 05:00 UTC

arXiv:2610.42116v1 Announce Type: new Abstract: In real-world deployments, large language model (LLM) systems are increasingly evolving from single-model invocation toward heterogeneous deployments in which different models are dynamically…

arXiv cs.AI 05:00 UTC

arXiv:2610.42096v1 Announce Type: new Abstract: World action models (WAMs) learn latent variables from observations and use these to both predict the evolution of the scene and to generate actions. The prediction has to be of the whole…

arXiv cs.AI 05:00 UTC

arXiv:2610.42066v1 Announce Type: new Abstract: AI answers require more than apparent correctness: reviewers need to inspect their recorded basis. We introduce the Knowledge Layer Model (KLM), an open, implementation-neutral specification…

arXiv cs.AI 05:00 UTC

arXiv:2610.41009v1 Announce Type: cross Abstract: Vision-language-action models (VLAs) map language instructions and visual observations to robot actions, showing strong capabilities across robotic manipulation tasks. However, in our…

arXiv cs.AI 05:00 UTC

arXiv:2610.41037v1 Announce Type: cross Abstract: Web overlays, such as cookie consent banners, newsletter prompts, and age gates, obstruct automated web crawlers and degrade downstream document representations. While proprietary…

arXiv cs.AI 05:00 UTC

arXiv:2607.18368v5 Announce Type: replace Abstract: Partially observable reinforcement learning requires deciding what to retain, retrieve, and forget over time. We introduce a neuro-symbolic meta-policy that learns which symbolic memory…

arXiv cs.AI 05:00 UTC

arXiv:2604.25907v3 Announce Type: replace-cross Abstract: The standard recipe for training reasoning models---SFT on rationale annotations followed by RLVR---relies on rationale annotations being available and clean. When no annotated…

arXiv cs.AI 05:00 UTC

arXiv:2610.41039v1 Announce Type: cross Abstract: As Mixture-of-Experts (MoE) large language models are increasingly deployed in distributed environments, from federated learning to edge-cloud systems, their architecture harbors an…

arXiv cs.AI 05:00 UTC

arXiv:2610.41074v1 Announce Type: cross Abstract: Low-precision training reduces computation but introduces additional rounding error. We theoretically study how stochastic rounding (SR) affects prediction risk as model size, training…

arXiv cs.AI 05:00 UTC

arXiv:2610.34829v2 Announce Type: replace Abstract: Single-document corpus poisoning under black-box access poses a severe threat to Retrieval-Augmented Generation (RAG). However, existing attacks predominantly rely on explicit factual…

arXiv cs.AI 05:00 UTC

arXiv:2610.41130v1 Announce Type: cross Abstract: Learned world models ship as software components whose model versions, preprocessing, and test harnesses change between releases. Work on model updates already counts per-sample…

arXiv cs.AI 05:00 UTC

arXiv:2610.41168v1 Announce Type: cross Abstract: Although large language models have shown strong capabilities in domain-specific code generation, their performance in robotics software development, especially for the Robot Operating…

arXiv cs.AI 05:00 UTC

arXiv:2503.13309v3 Announce Type: replace-cross Abstract: Despite advancements in Computer-Aided Diagnosis (CAD) systems, breast cancer remains one of the leading causes of cancer-related deaths among women worldwide. Recent breakthroughs…

arXiv cs.AI 05:00 UTC

arXiv:2605.06535v2 Announce Type: replace-cross Abstract: In recent years, open-source efforts like Senorita-2M have propelled video editing toward natural language instruction. However, current publicly available datasets predominantly…

arXiv cs.LG 05:00 UTC

arXiv:2506.13554v3 Announce Type: replace Abstract: We present a unified theoretical framework for analyzing the stability and consistency of Physics-Informed Neural Networks (PINNs), grounded in operator coercivity, variational…

arXiv cs.LG 05:00 UTC

arXiv:2610.42109v1 Announce Type: cross Abstract: Distributionally robust optimization (DRO) minimizes worst-case risk under distributional uncertainty, but its decisions can be suboptimal in terms of their excess risk, or regret. We study…

arXiv cs.AI 05:00 UTC

arXiv:2610.41172v1 Announce Type: cross Abstract: Retrieval-augmented generation is becoming agentic. A search harness writes queries, reads results, reformulates, and decides when to stop. This largely solves accuracy: a harness with only…

arXiv cs.AI 05:00 UTC

arXiv:2610.41304v1 Announce Type: cross Abstract: Particle image velocimetry (PIV) measures fluid motion from images of tracer particles, but its accuracy depends on choices throughout the estimation pipeline, including image…

arXiv cs.AI 05:00 UTC

arXiv:2505.22442v4 Announce Type: replace-cross Abstract: Offline RL (ORL) promises safe and sample-efficient deployment but existing methods rely on undocumented online interactions for hyperparameter tuning and lack reliable fully…

arXiv cs.AI 05:00 UTC

arXiv:2610.41322v1 Announce Type: cross Abstract: Generating small molecules to target protein-protein interactions (PPIs) remains one of the most challenging yet high-reward frontiers in drug discovery. Unlike traditional small-molecule…

arXiv cs.AI 05:00 UTC

arXiv:2610.41380v1 Announce Type: cross Abstract: Integrated sensing and communication (ISAC) is a core design principle for sixth-generation (6G) wireless systems, yet fixed antenna geometry limits its performance. Pinching antenna…

arXiv cs.AI 05:00 UTC

arXiv:2509.15688v2 Announce Type: replace-cross Abstract: Fine-grained visual classification benefits from high-resolution inputs to resolve subtle, localized cues, but at the cost of large compute and memory budgets. Inspired by human…

arXiv cs.AI 05:00 UTC

arXiv:2606.02242v2 Announce Type: replace-cross Abstract: Unifying image-based (I2I) and text-based (T2I) person re-identification (ReID) within a single model is complicated by a fundamental conflict between their training objectives…

arXiv cs.AI 05:00 UTC

arXiv:2610.41236v1 Announce Type: new Abstract: Sparse MoE mechanism significantly improves the trade-off between model capacity and inference efficiency. However, practical deployment still requires storing all experts, resulting in…

arXiv cs.AI 05:00 UTC

arXiv:2610.41264v1 Announce Type: new Abstract: Training language-model agents to play long-horizon strategy games is hard because deterministic feedback is sparse and often arrives only at the end of a game. In-game scores can offer dense…

arXiv cs.AI 05:00 UTC

arXiv:2610.41381v1 Announce Type: cross Abstract: In continuous spiking keyword spotting, temporally correlated acoustic frames can repeatedly trigger memory access and neural computation at a dense input interface, even when hidden…

arXiv cs.AI 05:00 UTC

arXiv:2610.41290v1 Announce Type: new Abstract: Learning-based approaches have increasingly been explored for SAT solving, with particular attention to improving branching decisions in conflict-driven clause learning (CDCL) solvers. These…

arXiv cs.AI 05:00 UTC

arXiv:2610.41245v1 Announce Type: new Abstract: Memory has become central to self-evolving LLM agents, which store and retrieve experience from past rollouts to guide future decisions. Existing systems represent this experience as…

arXiv cs.AI 05:00 UTC

arXiv:2610.41232v1 Announce Type: new Abstract: As Large Language Models (LLMs) become increasingly multilingual, distinguishing human-written from AI-generated text is essential for maintaining information integrity. However, research…

arXiv cs.AI 05:00 UTC

arXiv:2610.41183v1 Announce Type: new Abstract: LLM-guided program evolution can produce mathematical constructions, but validity, generality, and reproducibility require distinct evidence. We develop a researcher-in-the-loop discovery…

arXiv cs.AI 05:00 UTC

arXiv:2610.41169v1 Announce Type: new Abstract: The ability to integrate new information and deduce its consequences is central to learning, understanding, and reasoning in large language models (LLMs). Whether information arrives during…

arXiv cs.AI 05:00 UTC

arXiv:2610.41149v1 Announce Type: new Abstract: Agentic systems increasingly gate individual steps with a pointwise judge that scores a response against a criterion. Small language models are attractive for this role because they can be…

arXiv cs.AI 05:00 UTC

arXiv:2610.41129v1 Announce Type: new Abstract: AI agents increasingly carry out delegated tasks in parallel, with one person overseeing several of them at once. In practice, an agent often lacks information its task requires, such as an…

arXiv cs.AI 05:00 UTC

arXiv:2610.41107v1 Announce Type: new Abstract: Manifold learning methods like t-SNE and UMAP preserve local neighborhoods, but they do not preserve the ranking of the neighbors inside them. Two embeddings can return the same neighbor set…

arXiv cs.AI 05:00 UTC

arXiv:2610.41090v1 Announce Type: new Abstract: In text-to-SQL and AI-agent analytics, successful SQL execution does not establish semantic correctness under changing business definitions. We study revalidation of fixed queries after…

arXiv cs.AI 05:00 UTC

arXiv:2610.41085v1 Announce Type: new Abstract: Payroll-fraud screening requires finding hidden employee--payee links while distinguishing suspicious relationships from legitimate ones. We formulate this problem as inductive link…

arXiv cs.AI 05:00 UTC

arXiv:2610.41081v1 Announce Type: new Abstract: We study the so-called Minimum Red''Blue Separating Set Problem (MRBSSP) which arises from computational geometry. In particular, we propose a compact Set-Cover ILP formulation, safe…

arXiv cs.AI 05:00 UTC

arXiv:2610.41072v1 Announce Type: new Abstract: Security operations centers (SOCs) need to turn terabytes of alerts, logs and flows into trusted, affordable decisions to unveil hidden threats. We present LanG 2.0, an agentic system that…

arXiv cs.AI 05:00 UTC

arXiv:2610.41040v1 Announce Type: new Abstract: LLM agents serve natural-language requests through tools, data sources, and other agents, and even simple requests often need several interdependent calls. Typically, the LLM picks these…

arXiv cs.AI 05:00 UTC

arXiv:2610.41025v1 Announce Type: new Abstract: When a language-model agent fails a multi-step task, can another model find the decisive mistake and fix it with a single action? We introduce ReplayRCA, a benchmark that scores such…

arXiv cs.AI 05:00 UTC

arXiv:2610.40992v1 Announce Type: new Abstract: Managing multiple chronic conditions through diet requires satisfying disease- and medication-specific requirements jointly across meals and days while preserving established eating habits…

arXiv cs.AI 05:00 UTC

arXiv:2610.40970v1 Announce Type: new Abstract: Generative audio systems can produce large libraries of personalized bedtime content, but ordinary media objectives (retention, completion, and novelty) are poorly aligned with the goal of…

arXiv cs.AI 05:00 UTC

arXiv:2610.40940v1 Announce Type: new Abstract: Digital agents can now gather and organize information, reason over it, and use tools to complete complex tasks. Forecasting puts all of these abilities to the test: it requires integrating…

arXiv cs.AI 05:00 UTC

arXiv:2610.40939v1 Announce Type: new Abstract: Chain-of-thought (CoT) monitoring is a promising way to oversee increasingly capable agents. It succeeds in part because current models appear limited in their ability to control what they…

arXiv cs.AI 05:00 UTC

arXiv:2610.40938v1 Announce Type: new Abstract: Using customer conversations to train and evaluate financial language models raises confidentiality concerns because non-public financial facts can remain after direct identifiers are masked…

arXiv cs.AI 05:00 UTC

arXiv:2610.40917v1 Announce Type: new Abstract: Large language models act only through the infrastructure that surrounds them. That infrastructure, the harness, determines what a model can accomplish and how it fails, yet it has no settled…

arXiv cs.AI 05:00 UTC

arXiv:2610.40906v1 Announce Type: new Abstract: Recent scientific agents build knowledge recursively by retrieving evidence, constructing an intermediate state, and judging whether that state contains sufficient information. These agents…

arXiv cs.AI 05:00 UTC

arXiv:2610.40883v1 Announce Type: new Abstract: Voice-interactive large language models process both what a speaker says and how they say it, yet the two channels do not always agree. We ask whether that disagreement degrades speech…

arXiv cs.AI 05:00 UTC

arXiv:2610.40798v1 Announce Type: new Abstract: Web world models increasingly predict the consequences of candidate actions before agents act, yet a plausible predicted future does not necessarily support the correct safety decision. In a…

arXiv cs.AI 05:00 UTC

arXiv:2610.40758v1 Announce Type: new Abstract: Self-evolving agents rewrite a persistent state from feedback, but it is largely untested whether this lets them learn a rule nobody told them. An agent can also pass by writing down each…

arXiv cs.LG 05:00 UTC

arXiv:2610.38378v2 Announce Type: replace-cross Abstract: Generative recommendation (GR) reformulates item retrieval as autoregressive generation of discrete semantic identifiers (SIDs), offering a unified paradigm for personalized…

arXiv cs.LG 05:00 UTC

arXiv:2610.17117v2 Announce Type: replace-cross Abstract: Prior research into interpretability does not test whether a representation stays consistent under paraphrasing. Each pair of paraphrases can be thought of as a local witness that…

arXiv cs.LG 05:00 UTC

arXiv:2607.04449v2 Announce Type: replace-cross Abstract: Field-boundary accuracy depends on the model, sensor resolution, label discretization, and evaluation procedure. Measuring the contribution of sensor resolution therefore requires a…

arXiv cs.LG 05:00 UTC

arXiv:2606.07153v2 Announce Type: replace-cross Abstract: Physics-informed learning is increasingly used for partial differential equation (PDE)-governed inverse problems, but its reliability remains difficult to certify. This paper…

arXiv cs.LG 05:00 UTC

arXiv:2606.01764v2 Announce Type: replace-cross Abstract: We revisit the convergence guarantees of the Extragradient (EG) method for unconstrained biaffine min-max optimization. It is known that EG with a fixed stepsize achieves a…

arXiv cs.LG 05:00 UTC

arXiv:2605.28233v2 Announce Type: replace-cross Abstract: Fairness-accuracy trade-offs are a central concern in the deployment of fairness-aware machine learning methods. When sensitive attributes are unavailable at inference time-the so…

arXiv cs.LG 05:00 UTC

arXiv:2605.25592v3 Announce Type: replace-cross Abstract: We study optimal experimental design for multinomial logit (MNL) bandits, where an agent repeatedly selects a subset of $K$ items from a ground set of size $N$ and observes…

arXiv cs.LG 05:00 UTC

arXiv:2605.09342v2 Announce Type: replace-cross Abstract: Autonomous drone fleets have potential in medical supply delivery during disaster incident response. However, coordinating multiple drones in such settings introduces compounding…

arXiv cs.LG 05:00 UTC

arXiv:2603.16886v2 Announce Type: replace-cross Abstract: Reliable comparison of deep learning architectures for financial forecasting requires controlled evaluation that separates architectural effects from hyperparameter tuning, random…

arXiv cs.LG 05:00 UTC

arXiv:2603.08255v2 Announce Type: replace-cross Abstract: Tactile sensation is essential for contact-rich manipulation tasks. It provides direct feedback on object geometry, surface properties, and interaction forces, enhancing perception…

arXiv cs.LG 05:00 UTC

arXiv:2602.14020v2 Announce Type: replace-cross Abstract: Heavy-tailed data make covariance estimation sensitive to the clipping level: stronger clipping reduces variance but increases bias. We study how to choose this clipping level from…

arXiv cs.LG 05:00 UTC

arXiv:2602.01418v3 Announce Type: replace-cross Abstract: We propose Parabolic Positional Encoding (PaPE), a parabola-based positional encoding for vision modalities in attention-based architectures. Given a set of vision tokens-such as…

arXiv cs.LG 05:00 UTC

arXiv:2602.00076v2 Announce Type: replace-cross Abstract: We study neural networks (NNs) operating under hardware-induced memory faults and computational errors, which can degrade reliability in safety-critical applications. To address…

arXiv cs.LG 05:00 UTC

arXiv:2511.21715v3 Announce Type: replace-cross Abstract: This paper argues that dataset structure is important in image recognition tasks (among other tasks). Specifically, we focus on the nature and genesis of correlational structure in…

arXiv cs.LG 05:00 UTC

arXiv:2506.21741v3 Announce Type: replace-cross Abstract: Denoising diffusion probabilistic models (DDPMs) represent an entirely new class of generative AI methods that have yet to be fully explored. They use Langevin dynamics, represented…

arXiv cs.LG 05:00 UTC

arXiv:2408.17246v3 Announce Type: replace-cross Abstract: We consider the problem of learning maximal Lyapunov functions to compute the maximal region of attraction of nonlinear dynamical systems using neural networks. Two obstacles limit…

arXiv cs.LG 05:00 UTC

arXiv:2405.13302v2 Announce Type: replace-cross Abstract: We study how the Bauer--Jost--Liu curvature bound applies to hypergraph Ollivier--Ricci curvature (ORC) and how its accuracy depends on local mass and geometry. An explicit…

arXiv cs.AI 05:00 UTC

arXiv:2610.41308v1 Announce Type: new Abstract: Large language models increasingly generate structured outputs consumed by programs and downstream systems, where semantic correctness alone does not guarantee reliable execution. We…

arXiv cs.AI 05:00 UTC

arXiv:2610.41398v1 Announce Type: cross Abstract: We introduce GraphMP, a learning-based message-passing decoder for quantum low-density parity-check codes that transfers without fine-tuning across code families, code sizes, physical error…

arXiv cs.AI 05:00 UTC

arXiv:2512.12063v2 Announce Type: replace-cross Abstract: Domain models are central to software engineering, as they enable a shared understanding, guide implementation, and support automated analyses and model-driven development. Yet…

arXiv cs.AI 05:00 UTC

arXiv:2610.41358v1 Announce Type: new Abstract: Existing visual preference evaluation relies on costly human annotations or preference reward models, whereas online marketplace transaction prices are readily available and reflect real…

arXiv cs.AI 05:00 UTC

arXiv:2610.41368v1 Announce Type: new Abstract: Overly restrictive eligibility criteria contribute to widespread enrollment failure in clinical trials, yet published recommendations for how specific criteria should be broadened remain…

arXiv cs.AI 05:00 UTC

arXiv:2610.41420v1 Announce Type: cross Abstract: Gated Delta Networks (GDNs) offer linear-time sequence modeling, yet long-context prefill remains costly due to dense intra-chunk computation, recurrent-state updates, and intermediate…

arXiv cs.AI 05:00 UTC

arXiv:2610.41378v1 Announce Type: new Abstract: Parallel LLM agents tackle complex tasks by exploring multiple trajectories, yet useful discoveries remain isolated in individual execution histories. Existing sharing methods often select or…

arXiv cs.AI 05:00 UTC

arXiv:2610.41384v1 Announce Type: new Abstract: Evidence-grounded NLP systems often verify nested text regions sequentially, so verifying one region constrains others. We formalize this as Online Evidence-Containment Prediction (OECP)…

arXiv cs.AI 05:00 UTC

arXiv:2610.41482v1 Announce Type: cross Abstract: Recursive textual memory can carry earlier errors into later updates, yet task performance alone does not establish memory reliability. We define self-consistency bias as the persistence of…

arXiv cs.AI 05:00 UTC

arXiv:2512.14764v2 Announce Type: replace-cross Abstract: Modern operational systems ranging from logistics and cloud infrastructure to industrial IoT, are governed by complex, interdependent processes. Understanding how interventions…

arXiv cs.AI 05:00 UTC

arXiv:2607.19195v3 Announce Type: replace-cross Abstract: Using large deviations theory, we solve and obtain a general expression for the free energy functional for a broad class of associative memories, including dense associative…

arXiv cs.LG 05:00 UTC

arXiv:2511.02496v3 Announce Type: replace Abstract: We study latent geometry as an explicit component of representation quality in data-scarce learning. For an encoder (\phi), we define (Q_{\beta,\gamma}(\phi)=I(\phi(X);Y)-\beta\mathcal…

arXiv cs.AI 05:00 UTC

arXiv:2610.41397v1 Announce Type: new Abstract: Computer-use agents act through graphical interfaces, command lines, APIs, scripts, and Model Context Protocol servers. A policy decision can be interface-invariant while its executable…

arXiv cs.AI 05:00 UTC

arXiv:2610.41453v1 Announce Type: new Abstract: Tool timeouts can leave AI agents facing ambiguous tool outcomes and partial execution: an action may have failed, committed or remained in flight while other workflow steps have already…

arXiv cs.AI 05:00 UTC

arXiv:2610.41563v1 Announce Type: cross Abstract: Deploying 6G network slices across administrative domains requires an online decision for each arriving request: which domain hosts each virtual network function, how traffic crosses domain…

arXiv cs.AI 05:00 UTC

arXiv:2610.41511v1 Announce Type: new Abstract: Large language models (LLMs) can generate structured outputs, but existing constrained-decoding methods mainly enforce syntax and cannot directly enforce satisfaction of a formal…

arXiv cs.AI 05:00 UTC

arXiv:2610.41564v1 Announce Type: new Abstract: Language agents increasingly tackle long-horizon tasks in interactive environments, where actions often depend on latent preconditions. The core difficulty is not trajectory length but…

arXiv cs.AI 05:00 UTC

arXiv:2610.41661v1 Announce Type: cross Abstract: Activation steering is an inference-time technique used to control or modify the behavior, tone, or safety properties of a large language model (LLM) without retraining the model's weights…

arXiv cs.AI 05:00 UTC

arXiv:2601.06844v2 Announce Type: replace-cross Abstract: Understanding the structure of complex, nonstationary, high-dimensional time-evolving signals is a central challenge in scientific data analysis. In many domains, such as speech and…

arXiv cs.LG 05:00 UTC

arXiv:2512.21208v4 Announce Type: replace Abstract: We develop a finite-dimensional sensitivity framework for studying stability in learning systems whose states include representations, parameters, and update variables. The central object…

arXiv cs.AI 05:00 UTC

arXiv:2610.41596v1 Announce Type: new Abstract: Data Assimilation (DA) reconstructs the state of a spatio-temporal system by combining sensor observations with the output of a numerical model. Since model states are high-dimensional while…

arXiv cs.LG 05:00 UTC

arXiv:2512.22991v4 Announce Type: replace Abstract: Multimodal learning has become a prominent research area, with the potential of substantial performance gains by combining information across modalities. At the same time, model…

arXiv cs.LG 05:00 UTC

arXiv:2601.16812v2 Announce Type: replace Abstract: In many learning tasks, certain requirements on the processing of individual data samples can arguably be formalized, in an intelligible way, as strict constraints of the underlying…

arXiv cs.AI 05:00 UTC

arXiv:2610.41616v1 Announce Type: new Abstract: Large language models (LLMs) acquire extensive factual knowledge during pre-training, which may include sensitive, private, or copyrighted information, motivating the development of…