Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Look Up Say More Mashable Selects Mashable Voices Safety Net Creator Hub Versus Gift Ideas For Everyone On Your List Switch Off Trending Now In My Bag All Series Joseph Green is the Global Shopping Editor for Mashable. He covers VPNs, headphones, fitness gear, dating sites, streaming, and shopping events like Black Friday and Prime Day. All products featured here are independently selected by our editors and writers. If you buy something through links on our site, Mashable may earn an affiliate commission. Access this free live stream from anywhere in the world with ExpressVPN .

German drone maker Helsing has enlisted Japanese e-commerce and finance firm Rakuten to help finalize a deal to sell its unmanned systems to Japan's army, a spokesperson for Rakuten said on Monday. Japan's Ground Self-Defense Force is currently testing Helsing's HX-2 strike drone, as Tokyo looks to modernize its military and counter an increasingly assertive China. The field tests will run until the end of September, the spokesperson said. Helsing struck an "initial agreement" for the Japanese government to use its systems earlier this year, a spokesperson for the Munich-based company said, adding that the move was made possible by an unnamed "local brokerage partner." Japan's Ground Self-Defense Force did not immediately respond to a request for comment.

From branding irons to iris scans, the ancient business of proving who you are has never been stranger--or more lucrative. There are now more bots online than people. The challenge used to be telling who was a person from who was a bot; today, the challenge is telling good bots and bad bots apart--with humanity in the lurch. Three summers ago, Sam Altman, the C.E.O. of OpenAI, posted on X a fifteen-second video of dozens of people--mostly young, veiled in face masks, and wearing super-hip, white-soled sneakers--waiting in a long, snaky line in the traffic-light-lit but otherwise dusky dark of an unnamed city that appears to have been Nagoya, Japan. "Crazy lines around the world," Altman wrote. "One person getting verified every 8 seconds now." In the video, the camera pans to the front of the line, where a woman is kneeling on the sidewalk as if in prayer, staring into a gleaming silver orb about the size of a volleyball that has been propped up on a stand about the height of an end table. It looks like the eyeball of a giant robot Cyclops. She is bowing before an artificially intelligent machine, seeking a certification that she is human. That eyeball, which is known as the Orb, was developed by Tools for Humanity, a San Francisco-and Munich-based tech company that Altman founded in 2019 with a twenty-five-year-old German physicist named Alex Blania, who is now its C.E.O.

The ruling holds that a company that designs, trains, operates, and manages an AI system must assume legal liability for any damages caused by the responses it generates. A local court in Germany has issued a ruling that could reshape the operation of search engines and artificial-intelligence-based chatbots worldwide. The Munich Regional Court preliminarily ruled that Google is liable for a series of false statements generated by its AI Overviews feature, requiring the company to prevent the dissemination of erroneous or inaccurate claims through its search engine. The ruling stems from a case first reported by the Decoder, in which two publishers discovered that Google's AI-generated summaries linked them, in certain searches, to questionable business practices, scams, and subscription-related frauds, without any basis for doing so. Earlier this year, the affected companies sent the tech giant a cease-and-desist letter, according to the report.

Inverse optimization (IO) seeks to infer the parameters of a decision-maker's objective from observed context--action data. We study noiseless IO, where demonstrations are generated by a ground-truth objective. We provide a high-probability ${O}(\frac{d}{T})$ generalization bound for the induced action set, where $d$ is the number of unknown parameters and $T$ is the size of the training dataset. We strengthen these guarantees under additional conditions that ensure uniqueness of the chosen action, bringing our IO guarantees in line with best-arm identification results in the bandit literature. We further show that the ${O}(\frac{d}{T})$ rate is tight over all consistent estimators considered here, and extend the result to both instantaneous and cumulative regret. Notably, the resulting regret lower bound matches the corresponding upper bounds in the adversarial setting, indicating that the stochastic IO setting is effectively adversarial for the class of estimators studied here. Finally, we propose a parameter-free algorithm with lower per-iteration complexity than generic solvers. Experiments validate the predicted rates and illustrate the tightness of our bounds.

Drone incursions over U.S. military bases reveal critical vulnerabilities in civilian infrastructure, from airports to energy grids, that experts say remain dangerously exposed.

'She had big googly eyes and was wearing a red hat knitted by one of the careworkers' Emma the Social Robot by Paula Hornickel. 'She had big googly eyes and was wearing a red hat knitted by one of the careworkers' Emma the Social Robot by Paula Hornickel. 'The first resident that Emma - a social robot - was introduced to was called Peter. After that, Emma assumed they were all called Peter, which everyone found hilarious. O ne morning in July 2025, I arrived in the small, quiet town of Albershausen in south-west Germany.

Deep learning (DL) has become a cornerstone of modern machine learning (ML) praxis. We introduce the R package mlr3torch, which is an extensible DL framework for the mlr3 ecosystem. It is built upon the torch package, and simplifies the definition, training, and evaluation of neural networks for both tabular data and generic tensors (e.g., images) for classification and regression. The package implements predefined architectures, and torch models can easily be converted to mlr3 learners. It also allows users to define neural networks as graphs. This representation is based on the graph language defined in mlr3pipelines and allows users to define the entire modeling workflow, including preprocessing, data augmentation, and network architecture, in a single graph. Through its integration into the mlr3 ecosystem, the package allows for convenient resampling, benchmarking, preprocessing, and more. We explain the package's design and features and show how to customize and extend it to new problems. Furthermore, we demonstrate the package's capabilities using three use cases, namely hyperparameter tuning, fine-tuning, and defining architectures for multimodal data. Finally, we present some runtime benchmarks.

Selection bias arises when the probability that an observation enters a dataset depends on variables related to the quantities of interest, leading to systematic distortions in estimation and uncertainty quantification. For example, in epidemiological or survey settings, individuals with certain outcomes may be more likely to be included, resulting in biased prevalence estimates with potentially substantial downstream impact. Classical corrections, such as inverse-probability weighting or explicit likelihood-based models of the selection process, rely on tractable likelihoods, which limits their applicability in complex stochastic models with latent dynamics or high-dimensional structure. Simulation-based inference enables Bayesian analysis without tractable likelihoods but typically assumes missingness at random and thus fails when selection depends on unobserved outcomes or covariates. Here, we develop a bias-aware simulation-based inference framework that explicitly incorporates selection into neural posterior estimation. By embedding the selection mechanism directly into the generative simulator, the approach enables amortized Bayesian inference without requiring tractable likelihoods. This recasting of selection bias as part of the simulation process allows us to both obtain debiased estimates and explicitly test for the presence of bias. The framework integrates diagnostics to detect discrepancies between simulated and observed data and to assess posterior calibration. The method recovers well-calibrated posterior distributions across three statistical applications with diverse selection mechanisms, including settings in which likelihood-based approaches yield biased estimates. These results recast the correction of selection bias as a simulation problem and establish simulation-based inference as a practical and testable strategy for parameter estimation under selection bias.

Supervised machine learning assumes that labeled data provide accurate measurements of the concepts models are meant to learn. Yet in practice, human labeling introduces systematic variation arising from ambiguous items, divergent interpretations, and simple mistakes. Machine learning research commonly treats all disagreement as noise, which obscures these distinctions and limits our understanding of what models actually learn. This paper reframes annotation as a measurement process and introduces a statistical framework for decomposing labeling outcomes into interpretable sources of variation: instance difficulty, annotator bias, situational noise, and relational alignment. The framework extends classical measurement-error models to accommodate both shared and individualized notions of truth, reflecting traditional and human label variation interpretations of error, and provides a diagnostic for assessing which regime better characterizes a given task. Applying the proposed model to a multi-annotator natural language inference dataset, we find empirical evidence for all four theorized components and demonstrate the effectiveness of our approach. We conclude with implications for data-centric machine learning and outline how this approach can guide the development of a more systematic science of labeling.