Unfortunately, choosing the best OPE algorithm for each task and domain is still unclear.

Respiratory audio, such as coughing and breathing sounds, has predictive power for a wide range of healthcare applications, yet is currently under-explored. The main problem for those applications arises from the difficulty in collecting large labeled task-specific data for model development. Generalizable respiratory acoustic foundation models pretrained with unlabeled data would offer appealing advantages and possibly unlock this impasse. However, given the safety-critical nature of healthcare applications, it is pivotal to also ensure openness and replicability for any proposed foundation model solution. To this end, we introduce OPERA, an OPEn Respiratory Acoustic foundation model pretraining and benchmarking system, as the first approach answering this need. We curate large-scale respiratory audio datasets ($\sim$136K samples, over 400 hours), pretrain three pioneering foundation models, and build a benchmark consisting of 19 downstream respiratory health tasks for evaluation. Our pretrained models demonstrate superior performance (against existing acoustic models pretrained with general audio on 16 out of 19 tasks) and generalizability (to unseen datasets and new respiratory audio modalities). This highlights the great promise of respiratory acoustic foundation models and encourages more studies using OPERA as an open resource to accelerate research on respiratory audio for health.

Unfortunately, choosing the best OPE algorithm for each task and domain is still unclear.

Google has fixed a critical vulnerability in Chrome versions 139.0.7258.154/155 According to Google, the vulnerability has not yet been exploited for attacks in the wild. The manufacturers of other Chromium-based browsers are expected to follow suit in the coming days. In the Chrome Releases blog post, Krishna Govind presents the eliminated vulnerability (CVE-2025-9478), which is treated as if it were discovered by external security researchers, but Google Big Sleep is named as the discoverer of the vulnerability. This is an "AI" tool based on Gemini for detecting security vulnerabilities and it's designed to detect vulnerabilities on its own without human assistance.

Can large language models (LLMs) accurately simulate the next web action of a specific user? While LLMs have shown promising capabilities in generating ``believable'' human behaviors, evaluating their ability to mimic real user behaviors remains an open challenge, largely due to the lack of high-quality, publicly available datasets that capture both the observable actions and the internal reasoning of an actual human user. To address this gap, we introduce OPERA, a novel dataset of Observation, Persona, Rationale, and Action collected from real human participants during online shopping sessions. OPERA is the first public dataset that comprehensively captures: user personas, browser observations, fine-grained web actions, and self-reported just-in-time rationales. We developed both an online questionnaire and a custom browser plugin to gather this dataset with high fidelity. Using OPERA, we establish the first benchmark to evaluate how well current LLMs can predict a specific user's next action and rationale with a given persona and history. This dataset lays the groundwork for future research into LLM agents that aim to act as personalized digital twins for human.

After a period of beta testing, version 120 of the Opera browser is now being rolled out to the public. The biggest piece of news in this particular update is a new built-in translation feature with support for over 40 languages, and the browser is now based on Chromium 135. To protect privacy, the new Translator feature doesn't pass any information on to third parties, and the translation itself is processed using AI (in partnership with Lingvanex) on Opera's European-based servers. Other improvements in Opera 120 include improved password management, enhancements to Split Screen mode, refinements to Tab Islands, a new Miniplayer for videos, and VPN Pro. Lastly, Opera 120 includes a fix for a serious zero-day vulnerability (labeled CVE-2025-6554) in the V8 JavaScript engine.

Google has released Chrome 138, the latest edition of the browser that fixes several vulnerabilities, now in versions 138.0.7204.49/50 According to Google, none of the vulnerabilities have been exploited in the wild yet. On the "What's New" page in Chrome, Google advertises the ability to use tab groups on the go. With Chrome, you can now synchronize not only passwords but also tab groups between your desktop and mobile devices. Google is also raising awareness of Chrome's ability to search, select, and copy text in scanned PDF documents, but Google doesn't mention that this first requires text recognition (OCR) and whether this takes place locally or in the cloud.

Offline policy evaluation (OPE) allows us to evaluate and estimate a new sequential decision-making policy's performance by leveraging historical interaction data collected from other policies. Evaluating a new policy online without a confident estimate of its performance can lead to costly, unsafe, or hazardous outcomes, especially in education and healthcare. Several OPE estimators have been proposed in the last decade, many of which have hyperparameters and require training. Unfortunately, choosing the best OPE algorithm for each task and domain is still unclear. In this paper, we propose a new algorithm that adaptively blends a set of OPE estimators given a dataset without relying on an explicit selection using a statistical procedure.

Respiratory audio, such as coughing and breathing sounds, has predictive power for a wide range of healthcare applications, yet is currently under-explored. The main problem for those applications arises from the difficulty in collecting large labeled task-specific data for model development. Generalizable respiratory acoustic foundation models pretrained with unlabeled data would offer appealing advantages and possibly unlock this impasse. However, given the safety-critical nature of healthcare applications, it is pivotal to also ensure openness and replicability for any proposed foundation model solution. To this end, we introduce OPERA, an OPEn Respiratory Acoustic foundation model pretraining and benchmarking system, as the first approach answering this need. We curate large-scale respiratory audio datasets ( \sim 136K samples, over 400 hours), pretrain three pioneering foundation models, and build a benchmark consisting of 19 downstream respiratory health tasks for evaluation.

More than 30 performing arts leaders in the UK, including the bosses of the National Theatre, Opera North and the Royal Albert Hall, have joined the chorus of creative industry concern about the government's plans to let artificial intelligence companies use artists' work without permission. They also urged the government to support the "moral and economic rights" of the creative community in music, dance, drama and opera. The 35 signatories of the statement include the chief executives of the Sadler's Wells dance theatre, the Royal Shakespeare Company, the City of Birmingham Symphony Orchestra and the Leeds Playhouse. The performing arts bosses added that they embraced advances in technology and were "participants" in innovation, but stated the government's plans risked undermining their ability to participate in the development and deployment of AI. Critics of the opt out plan have described it as unfair and impractical.