Results for “models”

arXiv (research papers) Phuong Tuan Dat

HINT-WAM: Hindsight-Informed Training for World Action Models

World action models (WAMs) have shown promising generalization and data efficiency in robot manipulation by coupling action generation with predictions of future visual dynamics. Translating these capabilities into reliable task execution, however, remains a challenge for existing post-training methods. Offline supervised fine-tuning provides dense supervision on expert trajectories, but its flow-matching action expert is trained only on noised demonstration actions and…

Read at arXiv (research papers)

arXiv (research papers) Randall Balestriero SI

Counterfactual World Models: JEPAs That Reason About "What If"

World models operate in a world they cannot fully observe. When a hidden disturbance, such as a current pushing a boat, changes the outcome of an action, a counterfactual question follows: what would have happened under a different action, subject to the same disturbance? Its answer separates bad decisions from bad luck. We prove that world models trained on factual transitions cannot answer it on their own, whatever their data or capacity. For many physical systems, the…

Read at arXiv (research papers)

arXiv (research papers) Amr Elshazly Security

LLM-Guided Cyber-Physical Attack Path Identification in Smart Grids

Identifying high-impact attack paths in cyber-physical power systems is critical for improving grid resilience. Existing methods typically rely on either structural topology analysis or simulation-based cascading-failure models. Topology-based approaches capture connectivity but lack adversarial reasoning, while simulation-driven methods are primarily suited for offline analysis and are less effective for rapid evaluation of evolving attack scenarios. As a result…

Read at arXiv (research papers)

TechCrunch AI Sarah Perez SI

Sam Altman warns against a future where AI, not humans, calls the shots

Sam Altman and Greg Brockman pushed back against Silicon Valley's growing debate over an AI takeover, while acknowledging the challenges of keeping advanced models under human control.

Read at TechCrunch AI

arXiv (research papers) Hiba Akhaddar Tech policy

Sex Differences in the Detection of Parkinson's Disease from Speech

Parkinson's disease (PD) is the second most common neurodegenerative disease in the world, and 70\% to 75\% of PD patients suffer from speech disorders at some stage. Since PD affects men more frequently than women, previous publications tend to rely on unbalanced datasets that potentially introduce sex biases into trained models. This study investigates sex differences in PD detection from speech using deep learning and traditional machine learning approaches. We evaluated…

Read at arXiv (research papers)

arXiv (research papers) Tianhua Chen SI

How Generative Models Generate: A Concise Guide to the Generative AI Landscape

The rapidly expanding generative AI landscape brings together methods that differ in architecture, learning objective, mathematical representation and sampling procedure. A concise, selectively organised technical guide can provide an accessible starting point for developing a coherent understanding of this increasingly complex landscape. This article takes a generation-centred approach, connecting how models produce new observations with the modelling and learning…

Read at arXiv (research papers)

arXiv (research papers) Emad Aghajanzadeh

DRIFT: Detachable Reasoning for Runtime-Flexible and Efficient End-to-End Driving

The integration of vision-language models (VLMs) into end-to-end (E2E) autonomous driving has demonstrated strong potential by providing scene description and reasoning capabilities. However, VLMs are computationally expensive, and switching language reasoning on and off at runtime without hurting closed-loop performance has not been demonstrated, which hinders practical adoption. To address these issues, we propose DRIFT, an E2E autonomous driving framework that integrates…

Read at arXiv (research papers)

arXiv (research papers) Zihe Huang SI

F$^3$: What Makes Good Experience in World-Action Models?

Deployed world-action models (WAMs) can keep improving only if they make good use of their own experience. Existing methods supply retrieved rollouts or demonstrations to the policy as context, yet such experience often turns previously solved tasks into failures. What, then, makes good experience for a WAM? We show that experience is valuable only in use: its value depends not on whether the source rollout succeeded but on what is stored, when it is used, and where it…

Read at arXiv (research papers)

arXiv (research papers) Eeham Khan SI

When Does Continual Pre-Training Help? A Controlled Study of Dialectal Adaptation for Québec French

Continual pre-training (CPT) is widely used to adapt pretrained language models to dialectal and non-standard language varieties, yet it remains unclear when this additional training is necessary, what drives its benefits, and whether aggregate evaluation scores adequately capture its effects. In this paper, we present a controlled study of CPT for Québec French language understanding across nine LLMs spanning different families and scales. Using a budget-matched protocol of…

Read at arXiv (research papers)

The Quantum Insider Matt Swayne

Bradesco and evolutionQ Begin Quantum Communication Tests to Assess New Data Protection Models

Insider Brief PRESS RELEASE — A pioneering initiative in Brazil’s financial sector seeks to study, in a controlled environment, the feasibility of the technology and its potential to strengthen the security of information exchange among companies, financial institutions and startups in the coming years. São Paulo, September 2026 — Bradesco, through inovabra, has begun pioneering

Read at The Quantum Insider

arXiv (research papers) William John Atkinson Tech policy

Hi Gemini, Who Should I Vote For? The Benefits (and Problems) of Using LLMs As Voting Advice Applications

Large language models (LLMs) can provide voters information about political issues, party positions, and even advise on which party to vote for. However, no previous study explores whether off-the-shelf LLMs could be used as voting advice applications (VAAs) to improve voter-party congruence. In this article, we address this gap by examining how Gemini 3 Flash advises a sample of 6,022 German survey respondents on which party to vote for, based on their policy preferences…

Read at arXiv (research papers)

arXiv (research papers) William John Atkinson Tech policy

Hi Gemini, Who Should I Vote For? The Benefits (and Problems) of Using LLMs As Voting Advice Applications

Large language models (LLMs) can provide voters information about political issues, party positions, and even advise on which party to vote for. However, no previous study explores how well off-the-shelf LLMs function as voting advice applications (VAAs). In this article, we address this gap by examining how Gemini 3 Flash advises a sample of 6,022 German survey respondents on which party to vote for, based on their policy preferences and priorities alone. In brief, we find…

Read at arXiv (research papers)

arXiv (research papers) Hemen Taukoor Security

The Limit Is in the Input, Not the Model: State-Dependent Command Safety in Natural-Language Interfaces to Industrial Controllers

Language models judging commands to industrial controllers may not receive machine state, the equipment's current condition. The same command can require opposite safety decisions in different states, so an incorrect answer may reflect missing information rather than poor reasoning. We propose an answerability audit: does the model receive the information needed to decide? The audit records the model's inputs and pairs the same command with two states requiring opposite…

Read at arXiv (research papers)

arXiv (research papers) Gonçalo Fernandes Security

Gradient-Based Defensive Poisoning Against System Identification Attacks on Cyber-Physical Systems

AI-driven attacks learn the models of cyber-physical systems (CPS) by passively observing controller and plant signals, then use this knowledge to act covertly. Because covertness depends on the accuracy of the identified model, degrading that accuracy is a viable countermeasure. This work applies Adversarial Machine Learning as a defense, injecting small perturbations into the measurements the attacker observes to degrade an artificial neural network (ANN)-based system…

Read at arXiv (research papers)

The Decoder Maximilian Schreiner SI

Anthropic, OpenAI, and Google models readily marked medics, journalists, and surrendering soldiers as kamikaze drone targets

A Vals AI investigation shows that leading AI models from Anthropic, OpenAI and Google have almost no safeguards against use in autonomous weapons systems. In tests, the models readily flagged civilians as military targets, and a simple video game framing in the prompt was enough to override even belated refusals. The providers have promised fixes. The article Anthropic, OpenAI, and Google models readily marked medics, journalists, and surrendering soldiers as kamikaze drone…

Read at The Decoder

The Verge Rani Molla SI

Remember the frontier AI pause? Chip companies don’t either

This summer, OpenAI paused its largest planned frontier training run after its AI agents accidentally hacked Hugging Face and early evidence suggested that its next model could independently find and exploit security flaws. In September, Anthropic CEO Dario Amodei called for slowing advances in the most powerful AI models, warning that they could be used

Read at The Verge

TechPowerUp TheLostSwede SI

(PR) NVIDIA RTX Spark Compact Desktops Out Now

New NVIDIA RTX Spark compact desktops from ASUS, Dell and MSI are available today, bringing local AI agents, image and video generation, and model fine-tuning to a dedicated Windows PC. With up to 128 GB of unified memory and NVIDIA's AI software stack, these systems provide room for larger local models alongside RTX-accelerated creative and development tools. A dedicated desktop can keep agents and scheduled tasks running while a laptop travels or handles other work…

Read at TechPowerUp

TechPowerUp TheLostSwede SI

(PR) ASUS Announces Availability of the ProArt GR1X

ASUS today announced the availability of the ProArt GR1X mini PC. Built on the NVIDIA RTX Spark superchip, ProArt GR1X is a Windows PC designed as an always-on home for local AI agents. It is able to run large models on-device, keep prompts and project files private, and work with the agent and creative tools people already use. ProArt GR1X can be specified with 128 GB and 64 GB of unified memory. Built for agentic AI, powered by RTX Spark AI agents work best when they can…

Read at TechPowerUp

arXiv (research papers) Joshua Aftergood

Gutzwiller Quantum-Classical Embedding for Dynamics Simulations of Correlated Electron Systems

Simulating the real-time dynamics of correlated electron systems requires a nonperturbative treatment of local correlations over long times, a challenge for classical methods. We present the time-dependent Gutzwiller quantum-classical embedding framework, t-GQCE, for generic multi-orbital lattice models with time-dependent Hamiltonians. Within the time-dependent ghost Gutzwiller approximation, the lattice dynamics is mapped onto a noninteracting quasiparticle lattice…

Read at arXiv (research papers)

arXiv (research papers) Rong Zou

ALIVE: Bring Foundation Features to Life with Events

Visual foundation models provide powerful representations for diverse tasks, but their features become stale as visual content changes between image observations. Refreshing them requires new frames and repeated runs of a computationally expensive image encoder. We introduce ALIVE, an event-based method that incrementally updates native foundation model features between image observations. Given an anchor image, ALIVE uses subsequent events to transport and correct existing…

Read at arXiv (research papers)

9to5Mac Ryan Christoffel Gadgets

iPhone 18, iPhone Air 2 get new release timing updates

Apple only launched high-end ‘Pro’ and ‘Duo’ iPhones this fall, marking a departure from previous years when more affordable models debuted too. But iPhone 18 and iPhone Air 2 are still coming, and a new report offers an update on release timing.

Read at 9to5Mac

MacRumors Mitchel Broussard Gadgets

Amazon Offering $100 Off Nearly All M4 iPad Air Models

Amazon this week has multiple early Black Friday discounts on the M4 iPad Air, providing $100 off nearly every model. These remain the best prices we've seen in the wake of Apple's summer price hikes, and many deals are being matched at Best Buy. Note: MacRumors is an affiliate partner with Amazon. When you click a link and make a purchase, we may receive a small payment, which helps us keep the site running. We haven't tracked this many deals on the M4 iPad Air since Prime…

Read at MacRumors

The Decoder Manuel Uth SI

OpenAI opens its AI systems to four independent on-site auditors for the first time

OpenAI is giving four independent audit organizations, including Apollo Research and METR, direct access to unreleased models, source code, and training data for the first time. The embedded reviews are designed to uncover hidden misbehavior and safety gaps. OpenAI can redact or delay reports, so how independent this oversight actually turns out to be remains an open question. The article OpenAI opens its AI systems to four independent on-site auditors for the first time…

Read at The Decoder

The Decoder Jonathan Kemper SI

Hugging Face now routes model requests through OpenRouter for the best price or speed

Hugging Face now routes language model requests through OpenRouter, which automatically selects the cheapest or fastest provider. Bring-your-own API keys for language models are going away. According to OpenRouter, the new routing cut capacity errors by 46 percent. Both partners face multibillion-dollar acquisitions. The article Hugging Face now routes model requests through OpenRouter for the best price or speed appeared first on The Decoder.

Read at The Decoder

arXiv (research papers) Jiayou Zhu Security

RECLAIM: Robot-Vision-Informed Adversarial Patch Defense via Embodied Attention Feedback

Vision-language-action policies control embodied agents through a perception-action loop grounded in physical interaction, where adversarial patches can corrupt action predictions and thereby alter subsequent observations. To mitigate these threats, many existing defenses rely on policy adaptation or auxiliary models, but require additional training or inference. We introduce RECLAIM, a training-free, inference-time defense that uses a frozen VLA's own attention response to…

Read at arXiv (research papers)

arXiv (research papers) Sean Lim Security

Probing agent overeagerness in realistic settings with continuation audits

Frontier coding agents increasingly exhibit overeagerness, taking surprising, sometimes reckless actions beyond a user's mandate --- reported in recent system cards from frontier model developers. These cases are surfaced through internal deployment monitoring, which requires private traffic and is out of reach for open-weight models and independent evaluators. Automated alignment audits are the natural alternative, but capable models increasingly recognise the synthetic…

Read at arXiv (research papers)

arXiv (research papers) Jingni Huang

When to Refine: Selective Search for Frozen Visual World Models

Planning with a visual world model requires deciding both which actions to execute and how long to search for them. We introduce Selective CEM Refinement (SCR), a learned rule for deciding whether to continue cross-entropy method (CEM) search with a frozen LeWorldModel. After five it- erations, the gate measures how far the mean first action chunk has moved, how much the mean terminal cost of elite plans has fallen, and how widely their first chunks vary. It returns the…

Read at arXiv (research papers)

NVIDIA Blog Justin Walker SI

NVIDIA RTX Spark Compact Desktops Out Now

New NVIDIA RTX Spark compact desktops from ASUS, Dell and MSI are available today, bringing local AI agents, image and video generation, and model fine-tuning to a dedicated Windows PC. With up to 128GB of unified memory and NVIDIA’s AI software stack, these systems provide room for larger local models alongside RTX-accelerated creative and development

Read at NVIDIA Blog

9to5Mac Ben Lovejoy Gadgets

Leaker claims Apple developing four models of 20th anniversary iPhone

Leaker Ice Universe says Apple is currently developing four models of the 20th anniversary iPhone which we’ve been referring to as the iPhone 20. We’re expecting next year’s flagship models to be a radical new design with an all-glass appearance. Why four models rather than two is unclear, but it seems possible Apple is testing two versions of both the Pro and the Pro Max to determine just how ambitious it can be with the final product …

Read at 9to5Mac

Older stories