fourier-features

A library of scalable Bayesian generalised linear models with fancy features

Image classification using SVM, KNN, Bayes, Adaboost, Random Forest and CNN.Extracting features and reducting feature dimension using T-SNE, PCA, LDA.

Unofficial pytorch implementation of the paper "Learnable Fourier Features for Multi-Dimensional Spatial Positional Encoding", NeurIPS 2021.

Implementation of two phase field approaches for the surface reconstruction problem. One based of the Modica-Mortola theorem and the other based on Ambrosio-Tortorelli | Master Thesis

Incremental Sparse Spectrum Gaussian Process Regression

Sparse spectrum Gaussian process regression

AG-MAE: Anatomically Guided Spatio-Temporal Masked Auto-Encoder for Online Hand Gesture Recognition

Aplicación web para calcular y graficar series, integral, transformada y transformada discreta de Fourier, así como transformada de Laplace y resolver EDOs

Multi-Architecture Coupled Ensemble Physics-Informed Neural Networks (MACE-PINN) for solving coupled partial differential equations. Implements parallel subnetworks with Fourier embeddings and adaptive loss weighting for Gray-Scott and Ginzburg-Landau reaction-diffusion systems.

SCFGP: Sparsely Correlated Fourier Features Based Gaussian Process

Time series regression modeling on a dataset of supermarket sales across years, with the Darts library in Python.

Forecasting Store Sales for Improved Decision-Making Using Machine Learning for Time Series Data

Tensorflow 2.0 implementation of fourier feature mapping networks.

Neural Fields for Sea Surface Height Interpolation.

O E-COMPRESS é um sistema híbrido de compressão de arquivos que combina redes neurais supervisionadas com técnicas tradicionais como gzip e validação de integridade por CRC e ECC. Ele segmenta arquivos em blocos independentes, codifica cada um com Fourier Features e redes MLP, e reconstrói com alta fidelidade.

PyTorch implementation of a receiver query neural surrogate for surface seismogram emulation in heterogeneous 2D acoustic media. Includes Fourier encoded receiver coordinates, balanced sampling, a masked pseudo physics informed trace loss, and evaluation tools

A Modern Spin on Ptolemy's Geocentric Universe Model

Image regression experiments with random fourier feature mapping using PyTorch

Learned Nyquist: fundamental aliasing limits of implicit neural representations (IEEE ICASSP SPTM). Achievability + converse, Monte-Carlo-validated, real data both sides.

A Fourier Feature Network implementation for Regression tasks, mainly 2D Image Regression. Theoretical and mathematical background is heavily based on the "Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains" by M. Tancik, P. P. Srinivasan, B. Mildenhall et al.