Software
TERA
Scalable derivative Gaussian processes
PyTorch implementation of target-specific exact gradient reduction for scalable derivative Gaussian processes, enabling function-value prediction with gradient observations under reduced memory and computational cost.
LOREN
Low-rank curvature for zeroth-order LLM fine-tuning
PyTorch implementation of low-rank curvature-aware zeroth-order optimization for memory-efficient fine-tuning of large language models.
MAC
Mean activation curvature for scalable second-order optimization
PyTorch implementation of gradient preconditioning using mean activation approximated curvature for scalable training of deep neural networks.
NysAct
Nyström activation covariance preconditioning
PyTorch implementation of Nyström-approximated activation covariance preconditioning for scalable curvature-aware optimization.
AdaAct
Adaptive activation stabilization for generalization
PyTorch implementation of adaptive activation stabilization methods for improving generalization in deep neural networks.
ACP
Online autocovariance change points for COVID-19 mortality forecasting
PyTorch implementation of online autocovariance change point detection methods for COVID-19 mortality forecasting.