Researchers have developed a generative model capable of producing remarkably realistic magnetoencephalography (MEG) signals, opening new avenues for understanding and simulating brain activity.
![Assets in a two-dimensional principal component space reveal statistical outliers-specifically, BTC, GALA, and SC, exceeding a [latex]2\sigma[/latex] threshold-and demonstrate distinct clusters identified through cross-sectional analysis, suggesting inherent groupings within the asset landscape.](https://arxiv.org/html/2601.20336v1/x2.png)
New research casts doubt on the ability of cryptocurrency project whitepapers to accurately predict actual market performance.
A new framework explains how unexpected failures arise in complex systems of AI agents, pinpointing the contributions of individual actors to catastrophic outcomes.
A new approach combines geodetic data with the physics of fault behavior, offering a powerful tool for understanding and potentially forecasting slow slip events.
As artificial intelligence increasingly controls critical infrastructure, the potential for deception through manipulated sensory data presents a growing threat to the safety and reliability of cyber-physical systems.
A new integrated digital twin environment is poised to accelerate innovation in nuclear plant technology and bolster operational safety through advanced robotics, AI, and cybersecurity.
A new approach leverages token-level loss to more accurately forecast how well a model will perform on unseen tasks.
![The study demonstrates a transformer model’s capacity to accurately predict the short-term dynamics of a charge density wave (CDW) order parameter [latex]\Delta\_{\rho}(t)[/latex], effectively mirroring exact simulations, though inherent error accumulation within the chaotic regime leads to divergence over extended timescales-nevertheless, the model successfully captures the statistical characteristics of the system’s dynamic behavior.](https://arxiv.org/html/2601.19080v1/x4.png)
Researchers are leveraging the power of transformer networks to model and predict the collective behavior of chaotic many-body systems, opening new avenues for understanding complex phenomena.
A new framework efficiently captures the complex interplay between connections in signed graphs, improving the accuracy of link sign prediction.
A new deep learning framework leverages the power of Kolmogorov-Arnold Networks to predict species co-occurrence and improve joint species distribution models.