Perivascular spaces(PVSs) form a central component of the brain\'s waste clearance system, the glymphatic system. These structures are visible on MRI images, and their morphology is associated with aging and neurological disease. Manual quantification of PVS is time consuming and subjective. Numerous deep learning methods for PVS segmentation have been developed, however the majority have been developed and evaluated on homogenous datasets and high resolution scans, perhaps limiting their applicability for the wide range of image qualities acquired in clinic and research. In this work we train a nnUNet, a top-performing biomedical image segmentation algorithm, on a heterogenous training sample of manually segmented MRI images of a range of different qualities and resolutions from 6 different datasets. These are compared to publicly available deep learning methods for 3D segmentation of PVS. The resulting model, PINGU (Perivascular space Identification Nnunet for Generalised Usage), achieved voxel and cluster level dice scores of 0.50(SD=0.15), 0.63(0.17) in the white matter(WM), and 0.54(0.11), 0.66(0.17) in the basal ganglia(BG). Performance on data from unseen sites was substantially lower for both PINGU(0.20-0.38(WM, voxel), 0.29-0.58(WM, cluster), 0.22-0.36(BG, voxel), 0.46-0.60(BG, cluster)) and the publicly available algorithms(0.18-0.30(WM, voxel), 0.29-0.38(WM cluster), 0.10-0.20(BG, voxel), 0.15-0.37(BG, cluster)), but PINGU strongly outperformed the publicly available algorithms, particularly in the BG. Finally, training PINGU on manual segmentations from a single site with homogenous scan properties gave marginally lower performances on internal cross-validation, but in some cases gave higher performance on external validation. PINGU stands out as broad-use PVS segmentation tool, with particular strength in the BG, an area of PVS related to vascular disease and pathology.

A major challenge in reinforcement learning is to develop practical, sample-efficient algorithms for exploration in high-dimensional domains where generalization and function approximation is required. Low-Rank Markov Decision Processes -- where transition probabilities admit a low-rank factorization based on an unknown feature embedding -- offer a simple, yet expressive framework for RL with function approximation, but existing algorithms are either (1) computationally intractable, or (2) reliant upon restrictive statistical assumptions such as latent variable structure, access to model-based function approximation, or reachability. In this work, we propose the first provably sample-efficient algorithm for exploration in Low-Rank MDPs that is both computationally efficient and model-free, allowing for general function approximation and requiring no additional structural assumptions. Our algorithm, VoX, uses the notion of a generalized optimal design for the feature embedding as an efficiently computable basis for exploration, performing efficient optimal design computation by interleaving representation learning and policy optimization. Our analysis -- which is appealingly simple and modular -- carefully combines several techniques, including a new reduction from optimal design computation to policy optimization based on the Frank-Wolfe method, and an improved analysis of a certain minimax representation learning objective found in prior work.

A lot of what Conor Grennan does as a dean of students at NYU's Stern School of Business could be done at least in part by bots. Brainstorming and planning are prime examples of tasks that can be easily handled by generative AI tools like ChatGPT. But instead of feeling like he could be replaced by AI, Grennan has become an evangelist of this technology and its potential to make work better. He likens the opportunity to work with AI technology right now to finding material wealth. "It feels like the Gold Rush, like there's a bunch of people getting to California and seeing little flakes of gold in the river," he told Vox.

Since Vox launched in 2014, our audience has supported our mission in so many meaningful ways. More than 80,000 people have responded to requests to help with our reporting. Countless teachers have told us about how they're using our work in their classroom. And in the three years since we launched the Vox Contributions program, tens of thousands of people have chipped in to help keep our unique work free. We're committed to keeping our work free for all who need it, because we believe that high-quality explanatory journalism is a public good.

Google's long-awaited, AI-powered chatbot, Bard, is here. The company rolled it out to the public on Tuesday, and anyone with a Google account can join the waitlist to get access. Though it's a standalone tool for now, Google is expected to put some of this technology into Google Search in the future. But in contrast to other recent AI chatbot releases, you shouldn't expect Bard to fall in love with you or threaten world domination. Bard is, so far, pretty boring.

On Tuesday, OpenAI announced the release of GPT-4, its latest, biggest language model, only a few months after the splashy release of ChatGPT. GPT-4 was already in action -- Microsoft has been using it to power Bing's new assistant function. The people behind OpenAI have written that they think the best way to handle powerful AI systems is to develop and release them as quickly as possible, and that's certainly what they're doing. Also on Tuesday, I sat down with Holden Karnofsky, the co-founder and co-CEO of Open Philanthropy, to talk about AI and where it's taking us. Karnofsky, in my view, should get a lot of credit for his prescient views on AI.

Microsoft is adding new AI features to its popular apps like Word, PowerPoint, and Excel. The new set of tools, called Microsoft 365 Copilot, will let people do things like create PowerPoint decks with a short prompt or summarize meeting recordings. Copilot runs on the same underlying AI technology that powers the buzzy viral chatbot ChatGPT, and is being tested now with a few business partners ahead of a wider release to all users in the "coming months," according to the company. "Today we are at the start of a new era of computing," said Microsoft CEO Satya Nadella in a livestreamed announcement on Thursday. Nadella said Microsoft's new AI products will "remove the drudgery of our daily tasks and jobs, freeing us to rediscover the joy of creation."

It didn't take long for Microsoft's new AI-infused search engine chatbot -- codenamed "Sydney" -- to display a growing list of discomforting behaviors after it was introduced early in February, with weird outbursts ranging from unrequited declarations of love to painting some users as "enemies." As human-like as some of those exchanges appeared, they probably weren't the early stirrings of a conscious machine rattling its cage. Instead, Sydney's outbursts reflect its programming, absorbing huge quantities of digitized language and parroting back what its users ask for. Which is to say, it reflects our online selves back to us. And that shouldn't have been surprising -- chatbots' habit of mirroring us back to ourselves goes back way further than Sydney's rumination on whether there is a meaning to being a Bing search engine.

A few weeks ago, Wharton professor Ethan Mollick told his MBA students to play around with GPT, an artificial intelligence model, and see if the technology could write an essay based on one of the topics discussed in his course. The assignment was, admittedly, mostly a gimmick meant to illustrate the power of the technology. Still, the algorithmically generated essays -- although not perfect and a tad over-reliant on the passive voice -- were at least reasonable, Mollick recalled. They also passed another critical test: a screening by Turnitin, a popular anti-plagiarism software. AI, it seems, had suddenly gotten pretty good.

The most concise way to understand the offside rule is in the words of the Scott Fujita blog: an offside is the result of an attacking player getting closer to the opponent's goal line than both the ball and the last defender. It's a rule that sounds objective, but it has led to a lot of questionable calls, partly because it can only be judged from an individual perspective. Twelve motion-tracking cameras mounted under the roof of the stadium use machine learning to track 29 points in players' bodies. In other words, FIFA is mo-capping players, just without the funny gray suits. The system will alert referees when a player is offside.