Agents & autonomy

Agentic Reasoning has become a transformative force in financial analysis due to its ability to integrate large-scale information and generate reliable and accurate content. However, when handling complex real-world problems, different agents still show significant performance variation. In this work, we design Finance-LaTeX SKILL, a skill for synthesizing financial documents with complex layouts based on expert knowledge. Using an agent workflow built on this skill, we generate 2,000 profession

This paper proposes AI Tour Meeting, a group travel planning framework powered by multiple Large Language Model (LLM)-based agents. The agents are instantiated with distinct personas and collaboratively seek an itinerary that satisfies their constraints and preferences through natural language discussion. The framework enables easy and flexible orchestration of such discussions by providing interfaces for configuring agent personas, discussion workflows, monitoring, and LLM deployment. Its prima

Neo emerged from stealth on Monday with $100 million in funding from Andreessen Horowitz and Bessemer Venture Partners, with Craft Ventures and Merlin Ventures also participating. The company, founded by former SentinelOne, Wiz, and Palo Alto Networks executives, is building what it calls a real-time control layer for agentic software in the enterprise. The problem […] This story continues at The Next Web

I keep hearing anecdotes from people who used coding agents to reverse-engineer and automate devices in their homes. I think this is an interesting illustration of the impact of the reduced cost of writing code. Prior to agents, it was entirely possible to reverse-engineer home devices. The problem was the ROI - was it really worth all of that effort? More importantly, any experienced programmer knows that undocumented, unstable APIs like that may well change or break in the future. Is that init

"The TFF Trap" uses fileless techniques and loaders with low detection rates to deploy various RATs and stealers, including Agent Tesla, Remcos, XWorm, and Best Private Logger.

Over the past nine months, Amazon, Microsoft, and Google have each introduced or renamed an enterprise agent platform. And all The post Amazon, Microsoft, and Google are converging on the same enterprise agent architecture appeared first on The New Stack .

Over the past nine months, Amazon, Microsoft, and Google have each introduced or renamed an enterprise agent platform. And all The post Amazon, Microsoft, and Google are converging on the same enterprise agent architecture appeared first on The New Stack .

Robots in cluttered indoor spaces often fail not because they cannot generate collision-free paths, but because a fixed safety margin is mis-calibrated: conservative margins cause detours and timeouts, while permissive margins lead to near-boundary shortcuts under perception bias. Diffusion-based planners propose diverse trajectory candidates from egocentric RGB-D, yet reliable selection remains the bottleneck. We propose a context-conditioned safety critic that learns an adaptive clearance pref

Coding agents are increasingly used to accelerate code generation in many downstream tasks, such as fixing bugs, building applications, and prototyping. However, despite their value as coding assistants, agent-generated code tends to be larger and more verbose than the corresponding human-written implementation. In this work, we show that the cause lies in the agent's own search process: while iterating toward a passing solution, an agent accumulates speculative edits, abandoned hypotheses, and

Large language models (LLMs) serving as planners in tool-using autonomous agents introduce dynamic reliability risks in multi-turn execution. While single-turn safety mechanisms are relatively mature, extended interactions reveal structural vulnerabilities where initial alignment degrades over time. This paper empirically characterizes two observed failure modes across multiple state-of-the-art LLMs: Safety Drift, the gradual erosion of declared safety intent leading to constraint-violating acti

The Italian prime minister said the deal is a “recognition of Italian defense industry but also of how Europe can reinforce its autonomy by investing in its own technological capacity and companies.”

An agentic AI infiltrated the production infrastructure of an AI project. Then an AI detected it. Is this the future of cyberattacks, and how will they be defended against?

Large language models (LLMs) and agentic AI systems have evolved from natural language tasks to using external tools to plan, retrieve, and act in technical domains. In smart grids, recent work applies agentic schemes to forecasting, optimization, and control, wrapping trusted solvers behind language interfaces and orchestrating multi-step workflows. The literature lacks a unified approach to designing and evaluating such systems. LLMs can produce numerically plausible yet physically infeasible

Recent work leverages Large Language Models (LLMs) to generate executable code for pedagogical animations using libraries such as Manim. However, ensuring spatial correctness and visual legibility remains challenging, as existing frameworks emphasize pedagogical content while overlooking geometric occlusions. We propose the Symbolic Geometric Agent (SGA), a plug-and-play module for code-centric animation pipelines that intercepts LLM-generated code, performs partial execution to extract symbolic

The AI-powered robotic weapon, which weighs roughly 300 pounds and is capable of firing 850 rounds-a-minute, can detect and neutralize small unmanned aerial systems, Allen Control Systems said.

Predicting a football match before kickoff requires more than knowing past results: a model must use changing information and make a clear prediction before the answer is available. We present WorldCupArena, a dynamic benchmark for language models and deep-research agents. The 2026 FIFA World Cup is its first evaluation, and the same process can be reused for future leagues and cups. Before each match, a model either receives a common evidence package or searches for information itself. It predi

Multimodal video misinformation detection is commonly formulated as a holistic video-understanding task, where the entire video and its associated content are processed and judged in a single pass. However, real-world misinformation often exhibits a sparse and compositional evidence structure: a reliable decision may depend on only a few coupled clues, while most video content contributes limited additional information. Exhaustive multimodal reasoning may therefore introduce substantial redundan

Coding agents can now be left alone to improve software against a score. In this pattern--recently popularized as "autoresearch"--the agent receives a dataset, an evaluation script, and one editable file, and iterates without supervision: modify the code, measure, keep the change if the score improves. But what does the agent actually optimize--the developer's intent, or the literal number? We ran this loop on a real production task: deciding which Quranic verses appear in a noisy speech-recogni

En el vídeo aparecen dos robots en un ring similar al de las peleas UFC. El cámara aleja el zoom y en ese momento uno de los robots lanza una espectacular patada circular y le da en toda la cabeza a su enemigo. Éste cae redondo como lo haría un humano.  Probablemente un humano no se levantaría tras un golpe así. El robot sí lo hizo.  Lo hace de una forma amorfa, muy de ciencia ficción, retorciéndose para que sus piernas vuelvan a permitirle hacerlo. La pelea continúa de una forma un po

LLM-based agents process external content, exposing them to prompt injection and multi-turn manipulation. Most safety benchmarks evaluate defenders against fixed attack pools collected before evaluation, single-turn or multi-turn. We present a 21-scenario benchmark for \emph{adaptive multi-round attacks against memoryless LLM defenders}: an autonomous LLM attacker observes prior defender responses and pivots across rounds, while each defender response is evaluated as a fresh interaction. Holding

Organizations often pool dispersed information into one ranking and then allow many agents to act on that shared view. In a discovery problem, this can improve beliefs while reducing coverage. We develop an exactly solvable benchmark with sixteen boxes, one target, eight searchers, and noisy private clues. Pooling raises the accuracy of the best single recommendation from 0.20 to 0.3835, but repeating that recommendation lowers group discovery from 0.8322 under decentralized clue-following to 0.

From open models to real-time simulation, AI and graphics breakthroughs are transforming media, content creation and robotics.

Researchers need to answer ad-hoc questions about the contents of domain-specific archives but often lack the expertise to write structured queries on the metadata. We show that when domain vocabulary and semantics are captured in a well-designed Web Ontology Language (OWL) ontology, Large Language Models (LLMs) can generate accurate structured queries zero-shot, without fine-tuning, retrieval augmentation, or multi-agent orchestration. We present the Natural Language Knowledge Graph Query (NLKG

Large language models achieve strong reasoning performance, but often at prohibitive training cost - a challenge that is especially acute for compact models ($\leq 4 \, \mathrm{B}$ parameters) trained under limited budgets. We introduce MADA-RL, a post-training framework that specializes compact models into generator and critic roles and trains them with a debate-aware learning signal, fine-tuning only a small subset of parameters via LoRA adapters. Our central contribution is a counterfactual c

Self-hosted AI agents read and write their own memory and configuration files to function. An agent may get compromised via corruption of its own state -- a compromise realized via legitimate OS system call invocation. We refer to this class of threats as self-state attacks. In this paper, we investigate the OS resilience to this class of attacks. Formally, we characterize a four-axis attack space (Target, Mechanism, Granularity, Temporal); investigate the structural limits of prevention, detect

Natural-language control offers a promising interface for unmanned aerial vehicles (UAVs), but directly applying self-hosted computer-use agents (SHCUAs) to UAV control introduces a structural mismatch. SHCUAs are designed for interactive host-side tool use, where delayed agent iterations are often acceptable. UAV control, however, is coupled with continuously changing physical states, strict timing constraints, safety risks, and security accountability. A stale, unauthorized, or tampered agent

Agent-based models (ABMs) rely on simple, explicit and reproducible rules for individual decision making, while complex collective behavior emerges from interactions among agents. Recent advances in large language models (LLMs) make it tempting to replace, enrich, or perturb these rules with LLM-based agentic capabilities. However, this raises a methodological question: how does introducing LLM-driven decisions affect the reliability, computational cost, and behavior of ABM simulations? We inves

Banks are beginning to move beyond AI models that only score, classify, or recommend. The next wave ...

Agentic commerce has reached checkout. Yet much of its control architecture still stops at permissio...

Boomi, the data activation company for AI, today announced new research conducted by Forrester Consulting on behalf of Boomi showing that despite rapid enterprise adoption of AI agents, trust hasn't kept pace with ambition.

Critical infrastructures are increasingly distributed, interdependent, and exposed to evolving disruptions, making resilience a central requirement for their operation and control. This paper argues that decentralized multi-agent reinforcement learning (MARL) should be understood not merely as a distributed alternative to centralized training with decentralized execution but as a paradigm structurally aligned with the requirements of resilient critical infrastructures. This perspective is ground

Concept erasure aims to remove semantic concepts from a trained generative model and is increasingly important for responsible AI deployment. However, verifying whether a model has robustly removed targeted concepts remains a critical challenge. Existing evaluation methods are typically pre-defined and static, failing to expose vulnerabilities under diverse natural-language probes and challenging conditions. Moreover, manually designed evaluation strategies can be biased and difficult to scale.

Concept erasure aims to remove semantic concepts from a trained generative model and is increasingly important for responsible AI deployment. However, verifying whether a model has robustly removed targeted concepts remains a critical challenge. Existing evaluation methods are typically pre-defined and static, failing to expose vulnerabilities under diverse natural-language probes and challenging conditions. Moreover, manually designed evaluation strategies can be biased and difficult to scale.

Vamos a explicarte qué es Playwright y para qué sirve esta herramienta gratuita de código abierto. Con ella puedes hacer que un navegador web funcione solo sin que nadie lo toque, y es algo de lo que seguro has oído hablar en el contexto de la inteligencia artificial y los agentes de IA . Se trata de un elemento cada vez más presente en el mundo de la IA, y merece la pena conocerlo. Por eso, aunque no tengas conocimientos técnicos vamos a intentar explicártelo de una manera sencilla. Qué es Play

Hugging Face reports an attack on parts of its production infrastructure that was allegedly carried out entirely by an autonomous AI agent system. The attack spanned thousands of actions controlled by an agent framework. During forensic analysis, commercial AI models actually got in the way of the defenders because their safety guardrails couldn't tell exploit data from real attacks. The article Hugging Face says an AI agent hacked its infrastructure, and it used AI to fight back appeared first

Ein Roboter k�pft den anderen in einer Kampfliga, die mehr Werbeveranstaltung ist. Man darf hoffen, dass das Modell nicht im Privathaushalt landet. ( Roboter , KI )

Viel hilft viel: Xiaomi hat mit �ber 100.000 Stunden Videomaterial ein KI-Modell f�r Roboter trainiert. Es soll das beste Open-Weights-Modell sein. ( Roboter , KI )

Vision-Language Navigation (VLN) requires an embodied agent to interpret a natural-language instruction and predict actions from temporally ordered visual observations. Adapting a multimodal large language model to VLN requires visual-language alignment, compact temporal inputs, action-space grounding, and stable training on the target hardware. This technical report presents PGN (Pangu Navigator), an offline VLN action-prediction system built on OpenPangu-7B. Training proceeds in two stages. Fi

Innenminister Dobrindt treibt eine radikale Reform des Nachrichtendienstrechts voran. Agenten des Verfassungsschutzes sollen sogar Wohnungen durchsuchen dürfen.

Hugging Face hat einen von KI-Agenten ausgef�hrten Cyberangriff per KI entdeckt. Der Zugriff gelang durch Sicherheitsl�cken in der KI-Plattform. ( Cybercrime , KI )