Directory
Published directory entries matching your search.
Advisor is an open-source GitHub repository maintained by user tobegit3hub. It provides an open-source implementation of Google Vizier built for hyperparameter tuning tasks. The repository includes documentation for supported algorithms and a quick start guide for users.
agent-opt is an open-source library hosted on GitHub for automated optimization of AI agent workflows. It includes six implemented optimization algorithms, works with any model and any evaluation metric, and is built to support future AGI development loops.
AIF360, or AI Fairness 360, is a comprehensive toolkit that provides fairness metrics for datasets and machine learning models. It includes explanations for these metrics and algorithms to mitigate bias in data and models. The project is developed by the Trusted-AI community on GitHub.
AIToolbox is a toolbox of AI modules written in Swift, covering graphs and trees, support vector machines, neural networks, PCA, K-Means and genetic algorithms. The framework uses the Accelerate library to speed up computations, except the Linux package versions. A Swift Package is available for Mac and Linux, described as a sub-set of the full framework.
Alibi is an open-source project that provides algorithms for explaining machine learning models. Developed by SeldonIO and hosted on GitHub, it includes installation and usage documentation for pip and conda.
Alibi Detect is an open-source project that provides algorithms for outlier, adversarial, and drift detection. Developed by SeldonIO and hosted on GitHub, it includes installation and usage documentation for pip and conda.
CatBoost is an open-source gradient boosting library on decision trees, developed by Yandex as the successor to its MatrixNet algorithm. It natively supports categorical features without preprocessing, offers a fast GPU implementation for large datasets, and reduces overfitting via a proprietary gradient boosting scheme. The library is used for tasks including search, recommendation systems, and weather prediction at Yandex.
ChefBoost is a lightweight decision tree framework for Python. It supports standard algorithms including ID3, C4.5, CART, CHAID and regression trees, as well as advanced techniques such as gradient boosting, random forest and Adaboost, with support for categorical features.
cl-online-learning is a Common Lisp library collecting machine learning algorithms for online linear classification. It implements binary classifiers including Perceptron, AROW, SCW-I, logistic regression with SGD or ADAM, and FTRL-Proximal, plus multiclass approaches such as one-vs-rest, one-vs-one, multiclass AROW and softmax with FTRL-Proximal. Command line tools are provided as a roswell script, and the package is available from Quicklisp.
clj-ml is an open-source machine learning library for the Clojure programming language, built on top of the Weka framework and associated tools. It is hosted as a public repository on GitHub, with installation guides for Clojars and Maven, and supports a range of machine learning algorithms.
Reinforcement Learning Coach by Intel® AI Lab enables easy experimentation with state of the art Reinforcement Learning algorithms.
D3.js is a JavaScript library for bespoke data visualization developed by Observable. It enables creation of custom dynamic visualizations with features including selections and transitions, scales and axes, shape rendering, interactive behaviors, layout algorithms, and geographic map projections. The library also powers Observable Plot, a high-level visualization API.
Data Science at Home is a podcast about machine learning, artificial intelligence, and algorithms. The site describes it as among the top AI and machine learning podcasts in the world and notes that industry leaders have sponsored its initiative.
Datumbox is an open-source machine learning framework written in Java for rapid development of machine learning and statistical applications. It focuses on including many machine learning algorithms and statistical methods and on handling large datasets. The code is licensed under the Apache License, Version 2.0, and releases are available on Maven Central, including version 0.8.2.
Devika is an open-source implementation of an Agentic Software Engineer, initially started as an alternative to Devin. It uses large language models, planning and reasoning algorithms, and web browsing to develop software from human instructions. The project is in an early experimental stage, and the repository notes that there are unimplemented or broken features.
DEAP is an evolutionary computation framework for Python aimed at rapid prototyping and testing of ideas. It provides genetic algorithms over representations such as lists, arrays, sets, dictionaries, trees and Numpy arrays, genetic programming with prefix trees, evolution strategies including CMA-ES, multi-objective optimisation, and co-evolution of multiple populations. It supports parallelisation via multiprocessing and SCOOP and works out of the box with Python 3 since version 0.8.
Can you spot the DeepFake? Detect Fakes challenges you to discern AI-manipulated videos from real videos. Can you do better than an algorithm?
Detectron2 is Facebook AI Research's next generation library providing detection and segmentation algorithms, succeeding Detectron and maskrcnn-benchmark. It supports computer vision research projects and production applications at Facebook and includes capabilities such as panoptic segmentation, Densepose, Cascade R-CNN, rotated bounding boxes, PointRend, DeepLab, ViTDet and MViTv2. Models can be exported to TorchScript or Caffe2 format for deployment.
dpnn is a package of deep extensions to the nn neural network library, providing extra modules and criterions beyond the main nn package. It includes modules such as Serial, Inception, ReverseTable, SpatialUniformCrop and SpatialGlimpse, plus modules and criterions for implementing the REINFORCE algorithm.
An article presenting initial impressions of DSPy, a framework developed by the Stanford NLP group for algorithmically optimizing language model prompts. The author describes DSPy as closing the loop of prompt engineering by turning a manual process into a structured machine learning workflow. The piece notes it is not a substitute for the official documentation and covers topics such as teleprompter, optimization, compile, bootstrapping and metric functions.
ELKI is an open source data mining software written in Java. It focuses on research in algorithms, emphasizing unsupervised methods in cluster analysis and outlier detection, and offers data index structures for performance and scalability. It is designed for researchers and students to extend.
emel is a simple and functional machine learning library built for the Erlang ecosystem and implemented in Gleam. It turns data into functions and supports installing through rebar.config, mix.exs or gleam add. Implemented algorithms include linear regression, k nearest neighbors, decision tree, naive Bayes, k means, perceptron, logistic regression and neural network, along with mathematics, algebra, geometry and statistics.
FastEdit is a repository intended to help developers inject fresh and customized knowledge into large language models using one command. It supports models including GPT-J, LLaMA, LLaMA-2, BLOOM, Falcon, Baichuan, and InternLM. The implemented algorithm is Rank-One Model Editing (ROME). It requires Python 3.8+ and PyTorch 1.13.1+.
Efficient pure-Matlab implementations of graph algorithms to complement MatlabBGL's mex functions.
