March 2026

Probabilistic Retrofitting of Learned Simulators

⋅

Dominant approaches for modelling Partial Differential Equations (PDEs) rely on deterministic predictions, yet many physical systems of interest are inherently chaotic and uncertain. While training probabilistic models from scratch is possible, it is computationally expensive and fails to leverage the significant resources already invested in high-performing deterministic backbones. In this work, we adopt a training-efficient strategy to transform pre-trained deterministic models into probabilistic ones via retrofitting with a proper scoring rule: the Continuous Ranked Probability Score (CRPS). Crucially, […]

Ver mais

Liked Liked

BAED: a New Paradigm for Few-shot Graph Learning with Explanation in the Loop

⋅

The challenges of training and inference in few-shot environments persist in the area of graph representation learning. The quality and quantity of labels are often insufficient due to the extensive expert knowledge required to annotate graph data. In this context, Few-Shot Graph Learning (FSGL) approaches have been developed over the years. Through sophisticated neural architectures and customized training pipelines, these approaches enhance model adaptability to new label distributions. However, compromises in textcolor{black}{the model’s} robustness and interpretability can result […]

Ver mais

Liked Liked
February sponsors-only newsletter

February sponsors-only newsletter

⋅

I just sent the February edition of my sponsors-only monthly newsletter. If you are a sponsor (or if you start a sponsorship now) you can access it here. In this month’s newsletter: More OpenClaw, and Claws in general I started a not-quite-a-book about Agentic Engineering StrongDM, Showboat and Rodney Kākāpō breeding season Model releases What I’m using, February 2026 edition Here’s a copy of the January newsletter as a preview of what you’ll get. Pay $10/month to stay […]

Ver mais

Liked Liked

Dream2Learn: Structured Generative Dreaming for Continual Learning

⋅

Continual learning requires balancing plasticity and stability while mitigating catastrophic forgetting. Inspired by human dreaming as a mechanism for internal simulation and knowledge restructuring, we introduce Dream2Learn (D2L), a framework in which a model autonomously generates structured synthetic experiences from its own internal representations and uses them for self-improvement. Rather than reconstructing past data as in generative replay, D2L enables a classifier to create novel, semantically distinct dreamed classes that are coherent with its learned knowledge yet do […]

Ver mais

Liked Liked
Research roundup: Six cool science stories we almost missed

Research roundup: Six cool science stories we almost missed

⋅

It’s a regrettable reality that there is never enough time to cover all the interesting scientific stories we come across each month. So every month, we highlight a handful of the best stories that nearly slipped through the cracks. February’s list includes the revival of a forgotten battery design by Thomas Edison that could be ideal for renewable energy storage; a snap-on device to turn those boxers into “smart underwear” to measure how often we fart; and a […]

Ver mais

Liked Liked
Apple's new iPhone 17e has an A19 chip, MagSafe, and 256GB of storage for $599

Apple’s new iPhone 17e has an A19 chip, MagSafe, and 256GB of storage for $599

⋅

Apple’s biggest iPhone announcements usually happen in September, but for the second year in a row, the company is also introducing a new iPhone in March. The iPhone 17e is a new version of Apple’s basic no-frills iPhone, replacing last year’s iPhone 16e. The phone will be available to preorder on March 4 and will be available on March 11, starting at $599. The new iPhone includes an Apple A19 chip similar to the one in the more-expensive […]

Ver mais

Liked Liked

SEAR: Sample Efficient Action Chunking Reinforcement Learning

⋅

Action chunking can improve exploration and value estimation in long horizon reinforcement learning, but makes learning substantially harder since the critic must evaluate action sequences rather than single actions, greatly increasing approximation and data efficiency challenges. As a result, existing action chunking methods, primarily designed for the offline and offline-to-online settings, have not achieved strong performance in purely online reinforcement learning. We introduce SEAR, an off policy online reinforcement learning algorithm for action chunking. It exploits the temporal […]

Ver mais

Liked Liked
The Air-Gapped Chronicles: The UAE Sovereign Swap — Your US API Just Became a Regulatory Liability

The Air-Gapped Chronicles: The UAE Sovereign Swap — Your US API Just Became a Regulatory Liability

⋅

The Air-Gapped Chronicles: The UAE Sovereign Swap — Your US API Just Became a Regulatory Liability In 2026, the UAE isn’t consuming AI. It’s enforcing digital sovereignty. If your inference runs on US infrastructure, you’re one CBUAE audit away from shutdown. The $2.1M mistake: Routing patient data to US APIs violates CBUAE regulations. Sovereign gateway architecture keeps regulated data in UAE, routes only non-sensitive data globally. One passes audits. One gets penalized. Your healthcare AI processes 50,000 patient records daily […]

Ver mais

Liked Liked
It's almost a station wagon: The 2026 Subaru Trailseeker, driven

It’s almost a station wagon: The 2026 Subaru Trailseeker, driven

⋅

A new pair of EV siblings joins the Subaru lineup this year, each using a shared skateboard chassis developed in partnership with Toyota. Compared to the original Solterra and smaller Uncharted, the new Trailseeker bears more Subaru DNA despite riding on the same electric platform. And unlike the Solterra and Uncharted, the Trailseeker will be built for the American market alongside the Forester at the company’s Lafayette, Indiana, assembly line. Styling alone helps the Trailseeker look the most […]

Ver mais

Liked Liked
Former NASA chief turned ULA lobbyist seeks law to limit SpaceX funding

Former NASA chief turned ULA lobbyist seeks law to limit SpaceX funding

⋅

A former NASA administrator says he is “encouraged” that the US Congress is considering legislation to prevent NASA from spending more than 50 percent of its launch funding on any single provider. “America succeeds in space when American companies compete, innovate, and grow,” former NASA Administrator Jim Bridenstine wrote on LinkedIn. “I’m encouraged to see Congress taking meaningful steps to strengthen the industrial base that underpins both our civil and national security space missions.” Bridenstine commended the chairman […]

Ver mais

Liked Liked

Load More

↑

Digitado © 2026