Attention layers--which map a sequence of inputs to a sequence of outputs--are core building blocks of the Transformer architecture which has achieved significant breakthroughs in modern artificial intelligence. This paper presents a rigorous theoretical study on the learning and generalization of a single multi-head attention layer, with activation function replaced by ReLU, which have recently shown comparable performance with the original Softmax activation. We consider the random feature setting where the attention layer has a large number of heads, with randomly sampled frozen query and key matrices, and trainable value matrices. We show that such a random-feature attention layer can express a broad class of target functions that are permutation invariant to the key vectors. We further provide quantitative excess risk bounds for learning these target functions from finite samples, using random feature attention with finitely many heads. Our results feature several implications unique to the attention structure compared with existing random features theory for neural networks, such as (1) Advantages over standard fully connected random-feature models; (2) Concrete and natural classes of functions that can be learned efficiently by a random-feature attention layer. Additionally, we show that the sampling distribution of the query-key matrix (the product of the query and key matrix) matters--A biased Gaussian random matrix results in better sample complexities over the standard zero-mean counterpart for learning certain natural target functions.Experiments on simulated data corroborate our theoretical findings and further illustrate the interplay between the sample size and the complexity of the target function.

Boundary Representation (B-Rep) is the standard approach for modeling shapes in Computer-Aided Design(CAD).

Google's Chrome browser for MacOS and Windows is receiving an infusion of new Gemini-powered capabilities, including an AI browsing assistant contextually sensitized to a user's browsing activities. Google made the announcement this week at Google I/O 2025. Dubbed Gemini-in-Chrome, the feature will be available May 21 to Google AI Pro and Google AI Ultra subscribers in the US as well as Chrome Beta, Dev, and Canary users. The general idea behind Gemini-in-Chrome is to reorganize, aggregate, and then more sensibly redisplay the data found on one or more browser tabs while also embellishing the final output with additional but relevant Gemini-generated information. For example, during a pre-event press briefing attended by ZDNET, Google director of Chrome product management Charmaine D'Silva demonstrated how Gemini-in-Chrome could not only organize a head-to-head feature comparison chart of individual sleeping bags -- to which multiple Chrome tabs (one tab per sleeping bag) were pointing -- but could respond to text prompts about each bag's suitability to the expected temperatures for an upcoming camping trip in Maine.

Ahead of next week's Google I/O developer conference, where AI is expected to be the star of the show, Google on Tuesday unveiled a massive Gemini update, further expanding the AI assistant's availability across the Google ecosystem of devices. For starters, Google announced that Gemini will be coming to WearOS smartwatches later this year. This will allow users to take advantage of Gemini's conversational capabilities and talk to their watches naturally. The biggest advantage is that users can take advantage of hands-free assistance without the help of their phones. Some examples Google offers include setting reminders, getting assistance when your hands are covered in flour while cooking, or in the middle of a workout.

Microsoft is continuing its efforts to get straggling Windows 10 users to make the jump to Windows 11. Last week, the company published a new blog post (spotted by Windows Central) that shamelessly promotes its new Copilot PCs as the way forward for users still on older Windows 10 PCs that aren't eligible for the Windows 11 upgrade. Among other things, Microsoft is trying to entice you with brand-new AI features such as Recall, which takes periodic screenshots of your screen, analyzes it using AI, then indexes it so you can search through all your past PC activity at a later time. Another big benefit touted in the blog post is that Copilot PCs are purported to be up to five times faster than a five-year-old Windows device, with enhanced security features on top. Yet, while Copilot PCs are certainly novel and do offer plenty of benefits in the battery life and performance departments, they're still in need of some serious improvements.

The value of AI agents, systems that can carry out tasks for humans, is evident, with opportunities for productivity gains, especially for businesses. However, the performance of large language models (LLMs) can hinder the effective deployment of agents. Salesforce's AI Research seeks to address that issue. Also: 60% of AI agents work in IT departments - here's what they do every day On Thursday, Salesforce launched its inaugural Salesforce AI Research in Review report, highlighting the tech company's innovations, including new foundational developments and research papers from the past quarter. Salesforce hopes these pieces will help support the development of trustworthy and capable AI agents that can perform well in business environments.

Microsoft has officially set the expiration date on Windows 10 support, and once it hits, you'll stop receiving critical security updates. If you're still running the old system, this is your sign to upgrade to Windows 11 Pro. Windows 11 Pro brings a polished interface, faster multitasking, built-in AI with Microsoft Copilot, and upgraded security with features like Smart App Control and TPM 2.0. Gamers can take advantage of DirectX12 Ultimate for smoother, more immersive performance. This is a lifetime license for one compatible PC -- no subscriptions, no recurring costs.

Every time I get a press release about AI agents, I get a slightly queasy feeling. It is not quite as bad as that dizzy feeling I get every time someone insists on pitching me about vibe coding, nor is it the nails on a chalkboard feeling I get every time a PR rep sends me something with the word "convo" in it when asking for an interview or discussion with one of their clients. Even so, AI agents are overhyped, under-defined, possibly dangerous, and extremely limited; they could possibly spell the end of life as we know it. And yet everyone is all about agents. Microsoft did a series of announcements last week that promoted its extensive use of AI agents, not just for the enterprise, but for every Windows user.

Microsoft is celebrating its 50th anniversary, and the company is having some fun with it. The iconic Windows 95 logo was resurfaced, there is a themed version of Solitaire available, and Bill Gates even posted the source code for the company's first operating system, Altair Basic. Microsoft's Copilot is even getting some love. Actually, it would be more accurate to say that Microsoft has been showing Copilot a lot of love over the last few days. Announcements have been flying left and right, culminating in a livestream from Microsoft's global headquarters in Redmond, Washington, with even more information about current and upcoming Copilot features. Microsoft also had Copilot interview three Microsoft CEOs.

Microsoft turns the big 5-0 today (April 4), and the company is kicking off the celebration with a live Copilot event, rumored to include a suite of new announcements for the AI tool, as well as a nostalgic look back on the tech giant's storied history. Windows fans can expect an appearance from Copilot CEO Mustafa Suleyman -- and select deals on Microsoft products -- but the real star of the show is the iconic Windows 95 logo, back in all its glory to honor Microsoft's more humble beginnings. The live event will start at 12:30pm ET (9:30am PT). The Copilot event will stream live on the Microsoft Copilot YouTube page and the Microsoft website. Viewers can also catch birthday content, stories, and updates on Microsoft's 50th anniversary landing page.