Tag: Nature Machine Intelligence

Researchers at Colorado State University have developed flexible tensegrity robots inspired by elephant trunks that can shift shapes and mechanically ...

An international team of scientists has outlined a roadmap for developing ecological foundation models to integrate diverse environmental data and ...

A study in Nature Machine Intelligence reveals that advanced reasoning models consistently expend greater computational effort when processing counter-stereotypical associations.

Researchers introduced EVENET, a foundation model pretrained on 500 million simulated collision events that outperforms task-specific baselines across new-physics searches, ...

A physics-informed neural network called TrajectoryFlowNet predicts spatiotemporal flow fields and long-range particle trajectories from sparse measurements, validated on numerical ...

A systematic reusability study of the CORNETO framework reveals that multi-sample biological network inference operates in distinct robustness regimes governed ...

Researchers at ETH Zürich have shown that self-supervised pretraining on simulated detector data can teach a single neural encoder to ...

A new Nature Machine Intelligence perspective argues that the reasoning tokens produced by large language models should be understood as ...

A new framework called LACE uses large language models and complementary evolution to automatically design teams of specialist heuristics that ...

Recurrent neural networks trained on demanding multitask curricula spontaneously develop modular, brain-like architectures, showing that computational demands rather than wiring ...