In any future where humans and robots are working together, robots will need to be able to understand and respond to human emotions. Using a vision language model, researchers from Melbourne, Australia, developed a model to enable robots to respond differently depending on what emotions they are perceiving. https://spectrum.ieee.org/robot-emotions-visual-language-models?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

SpaceX’s orbital data centers promise abundant solar power and isolation from terrestrial risks. But when radiators the size of pickleball courts are required to keep a single AI rack at 60 °C, thermodynamics becomes a huge problem. https://spectrum.ieee.org/orbital-data-centers-heat?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

AI-enabled drug discovery is hard, but we’re getting better at it every day. Adrian Stecuła from Isomorphic Labs describes how it can be done. https://spectrum.ieee.org/isomorphic-labs-ai-drug-discovery?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

Videos intended to hold people in power accountable are being run through surveillance databases and may later be used to identify some of the same people who made the footage. This is the surveillance ouroboros. https://spectrum.ieee.org/unintended-consequences-video-surveillance?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

Scientists have developed a pacemaker that works without requiring surgery for battery replacement, using ultrasound waves. https://spectrum.ieee.org/pacemaker-and-ultrasound?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

Engineers have developed a process for training LLMs that uses 14 percent less energy without sacrificing speed. Their method involves adjusting the clock frequency of the GPU during computation. https://spectrum.ieee.org/llm-training-energy-saving-trick?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

A leading deep learning model for tracing glacier calving fronts can be adapted to new locations with only three pieces of information: one hand-labeled image per glacier, unlabeled summer reference images, and a map of the underlying rock. This could aid scientists who are tracking glacier loss. https://spectrum.ieee.org/tracking-glacier-melting-ai?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

To keep Curiosity doing science for 13 years, NASA has used some interesting fixes. Alexandra Holloway, the assistant team chief for engineering operations, shares the details. https://spectrum.ieee.org/curiosity-rover-jpl-mars-science?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

What if we put an incredibly powerful laser on the Moon? It could act as a time standard to synchronize lunar activity and as a spatial standard for creating GPS-like navigation for spacecraft. Jun Ye, a fellow at JILA, promises how it could be done. https://spectrum.ieee.org/lunar-lasers-optical-atomic-clock?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline

The National Science Foundation is issuing a total of US $1.5 billion to independent research organizations over the next ten years. This will support ambitious, high-risk science projects. https://spectrum.ieee.org/nsf-x-labs?utm_source=mastodon&utm_medium=social&utm_campaign=fedica-Mastodon-Daily-Pipeline