All papers

A genetic study of UK Biobank volunteers asks whether a napping habit actually changes the brain. The answer is a small yes for brain size, and a no for memory.

Implanted speech decoders work well on day one and degrade over weeks. A benchmark study tests which recalibration tricks actually help.

A fusion plasma can lose control in milliseconds. Can a learned controller keep it stable better than hand-tuned rules?

Speculative decoding promises faster text generation. This preprint measures what happens once real traffic, batching and memory limits enter the picture.

Adding more qubits usually adds more errors. This experiment shows a processor where the opposite happened, at least for now.

Does going to bed at wildly different times predict memory decline years later? A large cohort says yes, but the story is tangled.

A simulation study tests whether a simple curiosity bonus can help reinforcement learning agents find rewards that appear only once in a long while.

Thirty-eight planets between Earth and Neptune in size, and most seem to hide their atmospheres. What is getting in the way?

Models advertise windows of a million tokens. A new benchmark asks how much of that window they can really use when it counts.

A 12-week trial asked whether spaced recall practice helps older adults remember everyday information months later. The answer is encouraging, with caveats.