Nonlinear resonant interactions have long been theorized as the mechanism that moves energy across scales in ocean wind waves, but direct evidence of these processes in the open ocean has remained elusive. Here, the authors use a stereoscopic system to provide experimental evidence of resonant interactions in ocean wind waves, confirming theoretical predictions and offering further validation for third-generation ocean wave models.

The roughening transition in Z2 lattice gauge theory remains a complex phenomenon, particularly in understanding the behavior of electric flux strings. Here, the authors use numerical simulations with matrix product states to reveal the universal Lüscher correction and distinct entanglement entropy growth, offering insights into string dynamics and symmetry restoration near deconfinement.

Recent advances in large language models have expanded their capabilities beyond natural language processing, prompting exploration into their applications in high-energy physics. Here, the authors demonstrate that vision language models outperform traditional CNNs in classifying neutrino interactions, offering enhanced accuracy, robustness, and interpretability, thus paving the way for multimodal reasoning in experimental physics.

Artificial spin ices offer a way to explore emergent magnetic phenomena but are typically confined to flat geometries. Using Monte Carlo simulations, the authors determine the temperature-dependent behaviour of a 3D artificial spin ice based on the buckyball and show that it undergoes five crossovers on cooling, culminating in a ground state with a pair of topological defects at its poles.

Contextuality is a fundamental resource for quantum information processing, underpinning quantum entanglement. Here, the authors experimentally demonstrate the asymmetrical relationship between contextuality and entanglement using polarization-entangled photons, revealing the noninjective nature of their mapping and advancing methods to explore quantum properties.

Ecosystems in the Anthropocene exhibit complex dynamics that traditional ecological networks with pairwise interactions only partially capture. Here, the authors review hypergraphs as a framework for modeling higher-order interactions, showing how they clarify how ecological structures form and reorganize, with implications for conservation strategies including rewilding and habitat restoration aimed at sustaining biodiversity under rapid environmental change.

Efficient heat management in nanoscale electronics depends on the precise control of vibrational waves known as phonons. This research reveals that while periodic layered materials allow these waves to travel coherently, structural disorder triggers Anderson localization, trapping the waves and significantly reducing heat flow.

Astrophysical models underproduce the element Sr in intermediate neutron density environments (the so called i-process). The authors investigate the nuclear reactions that produce Sr and find that the new experimental reaction rates help to bridge the gap between models and observations.

Fluids lack positional order, which limits their ability to reshape objects and function as soft mechanical materials. Here, the authors show that a ferroelectric smectic-A liquid crystal can exert micronewton-scale forces and rotate and reshape embedded soft inclusions much more strongly than non-lamellar fluids.

Estimating nonlinear functions of quantum states is essential for quantum information processing but often computationally expensive. The quantum state function framework estimates a degree-n polynomial of a quantum state with \({{{\mathcal{O}}}}(n/{\varepsilon }^{2})\) copies without purified query access, enabling practical entropy, fidelity, and eigenvalue estimation.

Early-warning indicators such as increased variance, autocorrelation, and dimensional reduction are often interpreted as signs that complex systems in ecology, neuroscience, finance, or climate are approaching a tipping point at which a stable state loses stability or disappears. Here, the authors demonstrate, using brain dynamics during epileptic seizures as an example case, that the same statistical signatures can arise from pseudo-bifurcations in stochastic non-normal systems, showing that these widely used indicators are not uniquely diagnostic of tipping points.

Ultracold polar molecular gases provide a powerful platform for exploring quantum many-body physics with strong, long-range, and anisotropic interactions. Here, the authors explore quantum phases via beyond mean-field theory and path-integral Monte Carlo simulations, unveiling rich many-body behaviour.

AI applications in science often face limitations due to device-specific training and opaque “black-box” methods, hindering cross-system generalization. Here, the authors demonstrate a neural network trained on one tokamak’s turbulence data can predict edge localized mode onsets in another without retraining, showcasing potential for tokamak-agnostic AI systems.