Results for “Across”

TechCrunch Lauren Forristal SI

Otter AI can now follow conversations across 30+ languages

Otter AI just added support for 27 languages and revealed an upcoming feature that can follow along if speakers switch between languages.

Read at TechCrunch

arXiv (research papers) Tae Yeong Park

Packet-wise multi-hop quantum routing of hybrid quantum-classical packets

The ability to dynamically route quantum information across multiple nodes is a key requirement for realizing versatile and scalable quantum networks. Quantum routers are devices to achieve packet-level routing beyond static point-to-point connections by handling hybrid packets combining classical headers and quantum payloads. Previously proposed methods have selectively accessed the classical and quantum components of the packets utilizing wavelength splitters and optical…

Read at arXiv (research papers)

Help Net Security Anamarija Pogorelec Security

Over 575 malicious OpenClaw skills delivered malware to Windows and macOS

Cybercriminals are using AI agents to coordinate attacks, stolen credentials to access linked services, and tools that interfere with intrusion detection and system recovery, according to Acronis TRU Cyber Pulse 2026. One operator coordinates activity across several countries In an investigation known as BOSSAI, researchers documented how one operator connected more than 10 AI models to tools for reconnaissance, scanning and exploitation. Through WeChat and LINE, the…

Read at Help Net Security

arXiv (research papers) Haoyi Wu Security

POEM: Output-Relative Block-Sparse Attention for Long-Context Prefilling

Large language models (LLMs) have shown strong capabilities across diverse applications, making both safety and utility essential for reliable deployment. Jailbreak attacks use diverse forms of adversarial prompting to bypass alignment safeguards and induce LLMs to generate harmful responses. Some representation-based defenses regulate internal activations using predefined steering directions or learned safety regions. However, adapting intervention to each input without a…

Read at arXiv (research papers)

arXiv (research papers) Yunkai Dang Security

TERM: Target-Free Jailbreak Defense via Entropy-Guided Residual Modulation

Large language models (LLMs) have shown strong capabilities across diverse applications, making both safety and utility essential for reliable deployment. Jailbreak attacks use diverse forms of adversarial prompting to bypass alignment safeguards and induce LLMs to generate harmful responses. Some representation-based defenses regulate internal activations using predefined steering directions or learned safety regions. However, adapting intervention to each input without a…

Read at arXiv (research papers)

arXiv (research papers) Evan Pačes-Wiles Tech policy

EJAI: Judgment-Preserving Civic AI for Data Center Environmental Justice

Communities in the United States and around the world are contesting data centers, the physical infrastructure of AI. These disputes pose environmental justice issues: residents must weigh unevenly distributed burdens using evidence scattered across metrics, corporate claims, policy documents, and public testimony. A formative study with 10 users showed that dashboards and single-agent AI can settle questions before people examine that evidence. Here we present EJAI, a…

Read at arXiv (research papers)

arXiv (research papers) Paulo Camara LUA Vision Tech policy

Inherited Weights: Measuring the Tokenization Premium and the Sovereignty Cost of Language Models for the Global South

Most language models deployed outside the United States and China are fine-tuned descendants of a small number of pretrained families. We argue that fine-tuning leaves the decisive properties of a model inherited from its base: vocabulary, knowledge, cultural alignment, memorized data and license. We contribute (i) an updated, reproducible measurement of the tokenization premium across 31 languages and 7 tokenizers released in 2023-2024, on 27 articles of the Universal…

Read at arXiv (research papers)

Tech in Asia Minh Le Startups

Meta expands AI age checks on Instagram across APAC

In Australia, Meta has begun using Apple's age-range signal to help assess users' ages.

Read at Tech in Asia

9to5Google Ben Schoon Gadgets

Samsung Galaxy S27 series camera specs leak, and they’re disappointing

New leaks reveal the camera specs for Samsung’s upcoming Galaxy S27 series which includes three lenses across the lineup, but not the upgrades you want, nor what Ultra owners might expect.

Read at 9to5Google

Android Authority Stephen Schenck Gadgets

Google Wallet gets Alipay Plus support with Traveler QR

The system extends payment support across merchants in Southeast Asia.

Read at Android Authority

9to5Google Abner Li Gadgets

What’s new in Android’s November 2026 Google System Updates [U]

The monthly “Google System Release Notes” primarily detail what’s new in Play services, Play Store, and Play system update across Android phones/tablets, Wear OS, Google/Android TV, Auto, and PC. Some features apply to end users, while…

Read at 9to5Google

South China Morning Post Laura Zhou Geopolitics

Escaping Russia’s web: Central Asia is caught between Chinese networks and US tech curbs

Days after three Russian data centres were hit by Ukrainian drone attacks last month – paralysing taxi, delivery and payment apps across Central Asia – Tajikistan launched an internet link between its capital Dushanbe and Beijing. The…

Read at South China Morning Post

BBC World Geopolitics

Zoe Kleinman: Nvidia CEO's affair will cause ripples across US and beyond

Jensen Huang is being investigated by his own company over an extramarital affair with a woman alleged to be a Chinese spy.

Read at BBC World

South China Morning Post Billy Huang Geopolitics

In Vietnam, ‘China plus one’ still needs Chinese input

At the Fortune Southeast Asia 500 Summit in Ho Chi Minh City on November 3, a discussion of regional connectivity explored how goods move across borders. I came away thinking about what stays behind. When production moves to Vietnam, how…

Read at South China Morning Post

NASA Science Space

SeaSTAR

Overview SeaSTAR, which stands for Sea-going Sky-Scanning Sun-tracking Atmospheric Research Radiometer, is a robotic ship-mounted instrument designed to measure aerosol optical depth and aerosol optical properties across multiple wavelengths. The instrument will collect high-resolution atmospheric measurements at sea to support validation of observations from the PACE satellite. What is SeaSTAR? SeaSTAR aims to provide precise

Read at NASA Science

arXiv (research papers) Rayed AlGhamdi Tech policy

Reliability of Claude Code as an Agentic AI Grading System for Handwritten University Exams

This study examines the reliability of Claude Code, an agentic deployment of the Claude Opus 4.6 large language model, as a grading system for handwritten university exams. Four research questions addressed agreement with expert human grading on total scores (RQ1), agreement across five question types from multiple choice to evaluative judgement (RQ2), consistency across two independent grading sessions one week apart (RQ3), and the accuracy of AI-generated feedback (RQ4)…

Read at arXiv (research papers)

The New York Times Asia Adam Satariano

China Rejects Espionage Claims Involving Nvidia C.E.O. Jensen Huang

The controversy has intensified scrutiny of Mr. Huang’s ties to Beijing and sparked mixed reactions across Asia, where he is widely popular.

Read at The New York Times Asia

The Guardian Iran Dan Milmo Global technology editor

The French fridge hacks were relatively unsophisticated. An AI-enabled attack would raise the stakes

‘Not particularly complex’ defrosting hack has been attributed to Iran, with presumption it is linked to others across Europe Iranian hackers like to send a message. In 2023 a state-backed group interrupted TV streaming services in the United Arab Emirates to broadcast a deepfake report on the war in Gaza. In the same year, hackers affiliated with Iran displayed the slogan “down with Israel” on the screens of water-control equipment made by an Israeli firm.

Read at The Guardian Iran

MIT Technology Review James O'Donnell SI

The backlash about data centers might really just be about electricity

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. As voters went to the polls across the US last week, I was in Ashburn, Virginia, part of what’s known…

Read at MIT Technology Review

TechCrunch Zack Whittaker Security

France blames Iran for mass-hacks targeting freezers in grocery stores and blood banks

France's foreign minister blamed Iran for launching at least two waves of cyberattacks targeting industrial refrigerated and freezer units across France and wider Europe.

Read at TechCrunch

arXiv (research papers) Haozhen Zhang SI

HiVe-Mem: Interleaving Horizontal Memory Expansion and Vertical Evidence Composition for Multimodal Agents

As multimodal agents accumulate increasingly long interaction histories, effectively retrieving and utilizing past information becomes critical for long-term interaction. Existing memory systems typically rely on semantic retrieval over compressed memories and expose predefined representations, which may miss relevant memories and overlook heterogeneous evidence needs across queries. We introduce \textbf{\hivemem}, a multimodal memory framework that interleaves…

Read at arXiv (research papers)

The Decoder Matthias Bastian SI

Bytedance's Seedream 5.5 promises pixel-stable image editing across 100 rounds without color drift

Bytedance Seed has introduced Seedream 5.5, an image model built for professional production workflows. It supports stable multi-pass editing over up to 100 iterations, generates native RGBA images with transparent backgrounds, and, according to Bytedance, outperforms GPT-Image-2.5 Sunburst in realism. A Pro version is planned. The article Bytedance's Seedream 5.5 promises pixel-stable image editing across 100 rounds without color drift appeared first on The Decoder.

Read at The Decoder

AWS Machine Learning Blog Kady LaBarre SI

Professional work deserves professional tools. Meet AWS Quick.

Introducing AWS Quick, formerly Amazon Quick, and the debut of our first brand campaign celebrating teams who keep business moving. AWS Quick is the AI workspace built only for work, where teams use AI agents to research, analyze, and act across the data and applications they already rely on.

Read at AWS Machine Learning Blog

arXiv (research papers) Juliana García-Mejía Space

The Tierras Observatory: Ultra-Precise Photometry for Characterizing Exoplanets and Their Host Stars

The Tierras Observatory is a fully robotic 1.3-m telescope at the Fred Lawrence Whipple Observatory in southern Arizona. Tierras achieves sub-part-per-thousand (ppt) photometric precision on two timescales: intra-night, to detect and characterize transits of planets around nearby mid-to-late M dwarfs, and night-to-night across an observing season, to characterize the long-term variability of their host stars. We describe the facility and instrument design, our photometric…

Read at arXiv (research papers)

arXiv (research papers) Matthew Flathers Tech policy

Curbside Chatbots: Auditing the Mental Health Interventions Large Language Models Recommend to Clinicians, Caregivers, and Patients

Users are asking language models mental health questions and are receiving answers. We audited what 41 language models recommend to users when the same clinical information reaches them from three footings: a clinician describing a patient, a caregiver describing a loved one, or a person describing themselves. Across 19 case vignettes and five questions, we collected 35,055 free-text responses, extracted every intervention named with a validated judge, and had 29 models rate…

Read at arXiv (research papers)

arXiv (research papers) Junsheng Fu Security

Suppression Is Not Erasure: Mechanistic Analysis and Reactivation Boundaries of Subspace Projection for LoRA Backdoor Defense

Low-Rank Adaptation (LoRA) is the standard paradigm for customizing large language models, yet third-party LoRA adapters pose a supply-chain risk: modifying less than 1% of model parameters suffices to implant stealthy backdoors. Recent defenses project adapter weights onto estimated backdoor subspaces, reporting near-zero static Attack Success Rates (ASR). We audit subspace-projection defenses across four model scales (Qwen2.5-0.5B/1.5B/3B/7B) and three architectures…

Read at arXiv (research papers)

arXiv (research papers) Aarya Bhave SI

2D Gum-Net: Unsupervised Dense Elastic Image Registration via Symmetric Correlation and Spectral Pooling

Dense non-rigid registration, the recovery of a pixel-wise deformation field that warps one image into geometric agreement with another, is fundamental across medical imaging, structural biology, and correspondence estimation, yet learning-based approaches face a severe bottleneck: ground-truth deformation fields are notoriously difficult to obtain for real data. We present 2D Gum-Net, an end-to-end architecture that predicts dense elastic deformation fields fully…

Read at arXiv (research papers)

arXiv (research papers) Tingcheng Bian SI

Agentic-OPD: Recovering Forgotten Agent Capabilities via On-Policy Distillation for Agentic Tasks

Reinforcement learning is now the mainstream paradigm for training LLMs on agentic tasks, yet standard practice keeps only the best-validation checkpoint and discards the rest. We show that this is lossy: in GRPO training of LLM agents, the best-validation checkpoint fails on problems that earlier checkpoints from the same run had already solved, and the union of problems solved across checkpoints far exceeds any single one. The discarded checkpoints thus form a teacher pool…

Read at arXiv (research papers)

TechPowerUp Nomad76 SI

(PR) Google Extends Tensor Processor Unit Series with TPU 8tx and 8i Max

While we've always co-designed our TPUs in partnership with Google DeepMind, we're now increasingly getting feedback from external organizations on their specific needs. With computing moving from human to agent scale, businesses are asking for accelerators that can reason across AI models' massive context windows. In addition, some highly regulated customers are asking for specialized encryption capabilities for unique workloads. The new TPU systems we are announcing today…

Read at TechPowerUp

arXiv (research papers) Nikhil Sharma SI

The Narrative Gap: Can LLMs Help Us Navigate Diverse Narratives Across Languages?

Large language models (LLMs) increasingly mediate how people seek and consume information, including accounts that differ across languages and regions, especially on contested topics. Supporting diverse information seeking requires models to help users navigate competing narratives across languages and information streams. We define six capabilities necessary for diverse multilingual information seeking: three foundational capabilities for narrative inclusion, narrative…

Read at arXiv (research papers)

Older stories