Results for “model”

Tech.eu Tamara Djurickovic Startups

Cherry Ventures backs Tano with $4M to challenge the traditional influencer agency model

Creator advertisingplatform Tano has raised $4 million in a seed funding round led by Cherry Ventures, bringing its total funding to $5 million following earlier backingfrom Seedcamp and The Players F...

Read at Tech.eu

TechCrunch Tim Fernholz SI

Plurai wants to help AI agents work better using the same approach that helped get self-driving cars to roads

Like the end-to-end models trained to drive cars, AI agents are probabilistic systems operating in unpredictable environments.

Read at TechCrunch

The Decoder Jonathan Kemper SI

DeepL wants real-time translation to finally capture sarcasm and mood shifts

DeepL Voice's real-time translations are about to sound a lot more natural. The company's new Aure model carries over tone, emotion, emphasis, and speaking pace from the original speaker's voice into the target language. The article DeepL wants real-time translation to finally capture sarcasm and mood shifts appeared first on The Decoder.

Read at The Decoder

SecurityWeek Ionut Arghire Security

Fortinet Patches Nearly 100 Vulnerabilities

The massive drop comes after the company integrated frontier AI models into its product security testing.

Read at SecurityWeek

Digital Trends Shikhar Mehrotra

Apple wants to turn your Mac into a private AI server without making you touch the Terminal

A short-lived App Store listing shows Apple’s Claris building an easy way to run AI models across several linked Macs.

Read at Digital Trends

South China Morning Post Themis Qi Geopolitics

BYD apologises for wheel-rim error as China cracks down on car quality issues

BYD has apologised for an assembly error that caused some of its electric cars in China to be fitted with incorrect wheel rims, as Beijing ramps up scrutiny of quality issues in the country’s car industry. The world’s largest electric vehicle (EV) maker confirmed in a statement on Wednesday that a supplier had installed the front and rear wheel rims in reverse on some of its Tang model SUVs, offering free replacements to affected owners. The company said it had conducted…

Read at South China Morning Post

TechPowerUp AleksandarK

NVIDIA GeForce RTX 5080 24 GB Reportedly Cancelled, 16 GB Model Remains

NVIDIA's rumored GeForce RTX 5080 24 GB model has reportedly been put on hold. The company is struggling to secure enough 3 GB GDDR7 memory modules and is instead opting for the less expensive 2 GB modules for the standard 16 GB version. According to a report from Board Channels, NVIDIA has already informed its add-in card (AIC) partners to officially cancel the January 2027 launch plan for the 24 GB model, citing economic challenges. As we are currently experiencing one of…

Read at TechPowerUp

Help Net Security Anamarija Pogorelec Security

Microsoft researchers demonstrate hidden backdoor technique for AI agents

Microsoft researchers have demonstrated how small changes to an AI model can implant a hidden backdoor, causing an agent to execute an attacker’s command when a specific trigger appears. The technique, called Lazarus, requires no model retraining and allows the agent to behave normally during standard safety evaluations. The company has not observed threat actors weaponizing the technique, but says the research highlights a security concern for organizations that give AI…

Read at Help Net Security

YourStory Startups

Beyond the model: Why context, not intelligence, is the moat for durable AI

At TechSparks 2026, leaders from MongoDB, Pine Labs, and Spinny discussed why data, memory, workflows, and trust matter more than model access.

Read at YourStory

MacRumors Tim Hardwick Gadgets

Make Apple Watch Listen for Sounds Like the Doorbell

One of Apple's accessibility features on iPhone can listen for important sounds around you, such as a doorbell, siren or alarm, and it sends you a notification when it hears one. The feature has been extended to Apple's new Apple Watch Series 12 and Apple Watch Ultra 4 models, so the alert reaches your wrist even when your iPhone isn't with you. It's especially useful for people who are deaf or hard of hearing, but it has everyday uses too. If you wear noise-cancelling…

Read at MacRumors

arXiv (research papers) Kuang-Ting Huang

TQIR: Trellis Quantization with Integer Reconstruction for Bandwidth-Efficient Edge LLM Acceleration

Autoregressive large language model (LLM) inference is often limited by external-memory bandwidth, while existing ultra-low-bit trellis quantization reconstructs floating-point weights and leaves runtime Hadamard-transform and decoder costs unresolved. This paper presents TQIR, an algorithm--hardware co-design that decouples storage precision ($S$) from decoded-weight ($W$) and activation ($A$) compute precisions, enabling a better accuracy--hardware trade-off. TQIR performs…

Read at arXiv (research papers)

Help Net Security Industry News Security

Fingerprint trains specialized models to detect distinct fraud types

Fingerprint has launched Fraud Score, a predictive fraud signal trained directly on customers’ own fraud outcomes and powered by Fingerprint’s device intelligence and machine learning models. Fraud Score learns the fraud types at each organization to help them identify fraud more accurately without having to build and maintain their own models. Basic risk scoring can’t match the many types of fraud “Businesses today face a wider range of fraud threats, amplified by…

Read at Help Net Security

arXiv (research papers) Fengming Liu Tech policy

Where Language-Model Vocabulary Rose on Stack Exchange

Stack Overflow banned answers generated by ChatGPT five days after the model's release, and many other Stack Exchange sites announced similar rules. I measure the vocabulary of the answers posted afterwards, 3.8 million answers on 170 sites from 2019 to 2024. A fixed list of 56 words that language models overuse rose about 14 percent above its pre-2023 trend, while the highest of 1,000 frequency-matched lists of ordinary words rose about 7 percent. In a screen of 15 site…

Read at arXiv (research papers)

arXiv (research papers) Zhan Chen

AMARA: Compiler-Driven Accelerator Architecture Design for Graphical Probabilistic Sampling

Sampling methods are fundamental primitives for training and inference on probabilistic models. However, practical deployment of sampling algorithms such as Markov Chain Monte Carlo (MCMC) remains difficult on general-purpose processors. Existing hardware accelerators improve sampling efficiency, but they are typically specialized for a narrow class of models, graph topologies, or update schedules, making it difficult to support the diverse workloads used in real…

Read at arXiv (research papers)

Tech in Asia Naomi Li Gan Startups

Indosat subsidiary Lintasarta relaunches sovereign AI platform

Lintasarta said Deka LLM is powered by Indosat Intelligence Suite and gives users access to multiple AI models and supporting infrastructure.

Read at Tech in Asia

MarkTechPost Michal Sutter SI

MBZUAI’s IFM Releases PAN 2, an Open 29B World Model for Physical AI

The Institute of Foundation Models at MBZUAI released PAN 2, an Apache 2.0 world model that couples an LLM planner with a 29B diffusion transformer to simulate up to 20 seconds of action outcomes as video.

Read at MarkTechPost

BBC World Geopolitics

US warns Madagascar over growing military ties with Russia

US general warns the Sahel model-like security deal with Russia could lead to greater instability in Africa.

Read at BBC World

The Hacker News Security

Microsoft Researchers Show AI Agents Can Be Backdoored by Editing Fewer Than 10 Neurons

An AI model can be given a hidden backdoor by editing fewer than 10 of its rarely used neurons, without retraining the model, Microsoft researchers have shown. When a chosen trigger appears, the model outputs an attacker's command, which an agent using that model would run with the agent's own permissions. The attacker must be able to change the model's weights. The team tested the method on

Read at The Hacker News

Future Farming Geert Hekkert

Study: Small field robots could cut onion growing costs by 75%

Small, battery-electric field robots could substantially reduce production costs in onions and lettuce, according to a new study published in Nature Communications. But the impressive savings are based on theoretical modelling, not field trials. Actual costs and benefits may differ considerably under commercial farming conditions.

Read at Future Farming

Android Authority Hillary Keverenge Gadgets

Google rolls out Gemini 4 Carbon to AI Ultra subscribers

Google’s most capable AI model moves beyond its limited preview.

Read at Android Authority

Tech in Asia Naomi Li Gan Startups

Google DeepMind rolls out Gemini Robotics 2.5 to partners

Google said the new model improved task success rates, completion speed, and physical safety controls compared with earlier baselines.

Read at Tech in Asia

Inc42 Ankush Das Startups

Can Indian Startups Win The AI Inference Race?

Nations worldwide have spent the last few years obsessing over bigger models and the compute power needed to train them.…

Read at Inc42

The Register Security Security

OpenAI builds a kill switch for its research models

How much damage can a rogue agent do in 35 seconds? Asking cos that’s how long it might take to shut one down

Read at The Register Security

arXiv (research papers) Nadav Bojan Sellam SI

Moving Beyond the DRW Model: Learning Quasar Variability with Generative Models

The optical continuum variability of quasars is often described by a damped random walk (DRW) model, with characteristic damping timescales $τ_{\rm DRW}$ of hundreds of days and amplitudes $σ_{\rm DRW}$ of $\mathcal{O}(0.1)$ magnitudes. This model has been widely used as a null hypothesis to assess the significance of periodic signals in quasar light curves in the search for supermassive black hole binaries, the significance of extreme quasar variability like tidal…

Read at arXiv (research papers)

MarkTechPost Asif Razzaq SI

Fish Audio Releases Live 3, a Full-Duplex Voice Model That Reacts in 186 ms

Fish Audio released Live 3, a full-duplex voice model for voice agents. It listens while it speaks, separates “mm-hmm” from real interruptions, and can react in as little as 186 ms. English only at launch

Read at MarkTechPost

Tech in Asia Aiko Gao Ishida Startups

Lenovo launches AI token factory to tap China inference demand

The system routes complex tasks to larger AI models and simpler interactions to smaller ones.

Read at Tech in Asia

arXiv (research papers) Jiayang Li Security

Protecting Sensitive Knowledge Under LLM Unlearning: The Proxy Threat and a Superposition-Aware Defense

Machine unlearning lets the owner of large language models (LLMs) remove specific entities from a trained model without retraining from scratch. Existing machine unlearning methods focus on removing the requested entities while maintaining utility. However, we observe that the representations of distinct entities in an LLM may overlap in shared parameter subspaces, a phenomenon known as knowledge superposition. Removing one entity may then affect other entities. This exposes…

Read at arXiv (research papers)

arXiv (research papers) Arslan Bisharat Tech policy

Calibrate, Don't Retrain: Decomposing the Value of Annotator Disagreement in Cyberbullying Severity

Machine learning models are often trained on labeled data produced by multiple human annotators. When those annotators disagree, their labels are commonly collapsed into a single majority vote, which discards information about the underlying disagreement. We examine the cost of this practice for cyberbullying severity, a subjective judgment on which annotators disagree heavily. In the corpus we study, all 5 annotators assigned the same severity label to only 4.1% of…

Read at arXiv (research papers)

arXiv (research papers) Haoyang Zhang SI

EvoX: Rethinking the Agent Harness as a Runtime Policy

Large language model (LLM) agents increasingly interact with tools and environments, yet plausible model decisions do not always translate into successful execution due to tool-call mismatches, interrupted workflows, and unsupported completion claims. We present EvoX, a Rust-based, multi-model agent runtime guided by The Second Policy, which views the harness as an Operational Policy complementing the foundation model's Semantic Policy. EvoX combines model-facing guidance…

Read at arXiv (research papers)

Older stories