A new open-source neural network framework is presented, implemented from scratch without relying on pre-built deep learning modules. This framework serves as a pedagogical tool for understanding neural network mechanics and can be used as a baseline for educational purposes and future research.

Researchers developed Domain-aware Fourier Features (DaFFs) to enhance Physics-Informed Neural Networks (PINNs), improving performance and interpretability. DaFFs eliminate the need for explicit boundary condition loss terms and loss balancing schemes, simplifying PINN training and reducing computational cost.

Researchers investigate the effectiveness of in-context learning (ICL) in multi-agent decision-making, finding that the benefits of ICL in strategic environments may be due to statistical extrapolation rather than refined reasoning. This study provides a reusable framework for evaluating LLM reasoning in recursive belief tasks and introduces a diagnostic tool using rational expectations equilibrium (REE).

Researchers introduce Dual Randomized Smoothing, an improved technique for certifying neural network robustness against adversarial perturbations. This method allows for input-dependent noise variances, breaking through a limitation of the standard Randomized Smoothing formulation. Experiments show significant performance gains on CIFAR-10 and ImageNet datasets.

A new parameter pruning scheme for neural architectures is introduced, derived from the Fisher information metric. This method determines the optimal pruning level by computing the geodesic distance in the model space. The approach outperforms traditional magnitude pruning and local Fisher information pruning in various architectures and datasets, offering a state-of-the-art pruning methodology and a mathematically-motivated justification.