Displaying 1 to 18 from 18 results

celerite - Scalable 1D Gaussian Processes in C++, Python, and Julia

  •    C++

Read the documentation at: celerite.rtfd.io. The Julia implementation is being developed in a different repository: ericagol/celerite.jl. Issues related to that implementation should be opened there.

george - Fast and flexible Gaussian Process regression in Python

  •    Python

Fast and flexible Gaussian Process regression in Python. Read the documentation at: george.readthedocs.io.

revrand - A library of scalable Bayesian generalised linear models with fancy features

  •    Python

Note: we are not actively developing this library anymore, but we are still maintaining it. We recommend instead looking at Aboleth, which has similar functionality and is implemented on top of TensorFlow. revrand is a python (2 and 3) supervised machine learning library that contains implementations of various Bayesian linear and generalized linear models (i.e. Bayesian linear regression and Bayesian generalized linear regression).

bayesian-machine-learning - Notebooks related to Bayesian methods for machine learning

  •    Jupyter

This repository is a collection of notebooks covering various topics of Bayesian methods for machine learning. Gaussian processes. Introduction to Gaussian processes. Example implementations with plain NumPy/SciPy as well as with libraries scikit-learn and GPy.




AutoGP - Code for AutoGP

  •    Python

An implementation of the model described in AutoGP: Exploring the Capabilities and Limitations of Gaussian Process Models. The code was tested on Python 2.7 and TensorFlow 0.12.

k2sc - K2 systematics correction using Gaussian processes

  •    Jupyter

Python package for K2 systematics correction using Gaussian processes. where <filename> is either a MAST light curve filename, list of files, or a directory.

jlearn - Machine Learning Library, written in J

  •    J

WIP Machine learning library, written in J. Various algorithm implementations, including MLPClassifiers, MLPRegressors, Mixture Models, K-Means, KNN, RBF-Network, Self-organizing Maps. Models can be serialized to text files, with a mixture of text and binary packing. The size of the serialized file depends on the size of the model, but will probably range from 10 MB and upwards for NN models (including convnets and rec-nets).


aboleth - A bare-bones TensorFlow framework for Bayesian deep learning and Gaussian process approximation

  •    Python

A bare-bones TensorFlow framework for Bayesian deep learning and Gaussian process approximation [1] with stochastic gradient variational Bayes inference [2]. The purpose of Aboleth is to provide a set of high performance and light weight components for building Bayesian neural nets and approximate (deep) Gaussian process computational graphs. We aim for minimal abstraction over pure TensorFlow, so you can still assign parts of the computational graph to different hardware, use your own data feeds/queues, and manage your own sessions etc.

long-range-extrapolation - Capturing Structure Implicitly from Noisy Time-Series having Limited Data

  •    Jupyter

Scientific fields such as insider-threat detection and highway-safety planning often lack sufficient amounts of time-series data to estimate statistical models for the purpose of scientific discovery. Moreover, the available limited data are quite noisy. This presents a major challenge when estimating time-series models that are robust to overfitting and have well-calibrated uncertainty estimates. Most of the current literature in these fields involve visualizing the time-series for noticeable structure and hard coding them into pre-specified parametric functions. This approach is associated with two limitations. First, given that such trends may not be easily noticeable in small data, it is difficult to explicitly incorporate expressive structure into the models during formulation. Second, it is difficult to know a priori the most appropriate functional form to use. To address these limitations, a nonparametric Bayesian approach was proposed to implicitly capture hidden structure from time series having limited data. The proposed model, a Gaussian process with a spectral mixture kernel, precludes the need to pre-specify a functional form and hard code trends, is robust to overfitting and has well-calibrated uncertainty estimates. Bayesian modeling was adopted to account for uncertainty. Citation for the corresponding paper is as follows.

Miscellaneous-R-Code - Code that might be useful to others for learning/demonstration purposes.

  •    R

This is a place for miscellaneous R and other code I've put together for clients, co-workers or myself for learning and demonstration purposes. The attempt is made to put together some well-commented and/or conceptually clear code from scratch, though most functionality is readily available in any number of well-developed R packages. Typically, examples are provided using such packages for comparison of results. I would say most of these are geared toward intermediate to advanced folks that want to dig a little deeper into the models and underlying algorithms. More recently, if it gets more involved, I usually just create a document of some kind rather than a standard *.R file, so you might check out the docs repo as well.

GeoStats.jl - An extensible framework for high-performance geostatistics in Julia

  •    Julia

GaussianProcesses.jl — Gaussian processes (the method) and Simple Kriging are essentially two names for the same concept. The derivation of Kriging estimators, however; does not require distributional assumptions. It is a beautiful coincidence that for multivariate Gaussian distributions, Simple Kriging gives the conditional expectation. Matheron and other important geostatisticians have generalized Gaussian processes to more general random fields with locally-varying mean and for situations where the mean is unknown. GeoStats.jl includes Gaussian processes as a special case as well as other more practical Kriging variants, see the Gaussian processes example. MLKernels.jl — Spatial structure can be represented in many different forms: covariance, variogram, correlogram, etc. Variograms are more general than covariance kernels according to the intrinsically stationary property. This means that there are variogram models with no covariance counterpart. Furthermore, empirical variograms can be easily estimated from the data (in various directions) with an efficient procedure. GeoStats.jl treats variograms as first-class objects, see the Variogram modeling example.

go-bayesopt - A library for doing Bayesian Optimization using Gaussian Processes (blackbox optimizer) in Go/Golang

  •    Go

A library for doing Bayesian Optimization using Gaussian Processes (blackbox optimizer) in Go/Golang. This project is under active development, if you find a bug, or anything that needs correction, please let me know.

keras-gp - Keras + Gaussian Processes: Learning scalable deep and recurrent kernels.

  •    Python

KGP is compatible with: Python 2.7-3.5. In particular, this package implements the method described in our paper: Learning Scalable Deep Kernels with Recurrent Structure Maruan Al-Shedivat, Andrew Gordon Wilson, Yunus Saatchi, Zhiting Hu, Eric P. Xing arXiv:1610.08936, 2016.

gpr - Library for doing GPR (Gaussian Process Regression) in OCaml

  •    OCaml

Please refer to the GPR manual for further details and to the online API documentation as programming reference. Please submit bugs reports, feature requests, contributions and similar to the GitHub issue tracker.

flu-sequence-predictor - An experimental deep learning & genotype network-based system for predicting new influenza protein sequences

  •    Jupyter

An experimental deep learning & genotype network-based system for predicting new influenza protein sequences. Flu Forecaster was first fully developed during my time as an Insight Health Data Fellow. The projected business use case is to predict what future strains of flu will look like, which would thus help inform the pre-emptive development of vaccines. No longer would we have to select currently-circulating strains; instead, we could forecast what strains would look like 6 months down the road, pre-emptively synthesize them (using synthetic biology methods), and rapidly scale up production of the ones used for the flu shot vaccine.

Stheno.jl - Probabilistic Programming with Gaussian processes in Julia

  •    Julia

Stheno is designed to make doing non-standard things with Gaussian processes straightforward. It has an intuitive modeling syntax, is inherently able to handle both multi-input and multi-output problems, trivially supports interdomain pseudo-point approximations, and has some support for structure-exploiting algebra. First, a note for statistics / ML people who aren't too familiar with Julia: the first execution of the examples below will take a while as Julia has to compile the code. On subsequent runs (e.g. if you were repeatedly evaluating the logpdf for kernel parameter learning) it will progress much faster.

DynaML - Scala Library/REPL for Machine Learning Research

  •    Scala

DynaML is a Scala & JVM Machine Learning toolbox for research, education & industry. Interactive. Don't want to create Maven/sbt project skeletons every time you want to try out ideas? Create and execute scala worksheets in the DynaML shell. DynaML comes packaged with a customized version of the Ammonite REPL, with auto-complete, file operations and scripting capabilities.