About: A recommender systems research framework aimed at modeling non-stationary environments.

Initial Announcement on mloss.org.

About: Aika is an open source text mining engine. It can automatically extract and annotate semantic information in text. In case this information is ambiguous, Aika will generate several hypothetical interpretations about the meaning of this text and retrieve the most likely one.

Aika Version 0.17 2018-05-14

• Introduction of synapse relations. Previously the relation between synapses was implicitly modeled through word positions (RIDs). Now it is possible to explicitly model relations like: The end position of input activation 1 equals to the begin position of input activation 2. Two types of relations are currently supported, range relations and instance relations. Range relations compare the input activation range of a given synapse with that of the linked synapse. Instance relations also compare the input activations of two synapses, but instead of the ranges the dependency relations of these activations are compared.
• Removed the norm term from the interpretation objective function.
• Introduction of an optional distance function to synapses. It allows to model a weakening signal depending on the distance of the activation ranges.
• Example implementation of a context free grammar.
• Example implementation for co-reference resolution.
• Work on an syllable identification experiment based on the meta network implementation.

Aika Version 0.15 2018-03-16

• Simplified interpretation handling by removing the InterpretationNode class and moving the remaining logic to the Activation class.
• Moved the activation linking and activation selection code to separate classes.
• Ongoing work on the training algorithms.

About: Armadillo is a high quality C++ library for linear algebra & scientific computing, aiming towards a good balance between speed and ease of use. The function syntax is deliberately similar to MATLAB. Useful for algorithm development directly in C++, or quick conversion of research code into production environments (eg. software & hardware products).

• faster handling of sparse matrices by kron() and repmat()
• faster transpose of sparse matrices
• faster element access in sparse matrices
• faster row iterators for sparse matrices
• faster handling of compound expressions by trace()
• more efficient handling of aliasing in submatrix views
• expanded normalise() to handle sparse matrices
• expanded .transform() and .for_each() to handle sparse matrices
• added reverse() for reversing order of elements
• added repelem() for replicating elements
• added roots() for finding the roots of a polynomial

About: Kernel-based Learning Platform (KeLP) is Java framework that supports the implementation of kernel-based learning algorithms, as well as an agile definition of kernel functions over generic data representation, e.g. vectorial data or discrete structures. The framework has been designed to decouple kernel functions and learning algorithms, through the definition of specific interfaces. Once a new kernel function has been implemented, it can be automatically adopted in all the available kernel-machine algorithms. KeLP includes different Online and Batch Learning algorithms for Classification, Regression and Clustering, as well as several Kernel functions, ranging from vector-based to structural kernels. It allows to build complex kernel machine based systems, leveraging on JSON/XML interfaces to instantiate prediction models without writing a single line of code.

In addition to minor improvements and bug fixes, this release includes:

• The possibility to generate the Compositional GRCT and the Compositional LCT data structures in kelp-input-generator.

• New metrics for evaluating Classification Tasks.

• New Tutorial and Unit Tests.

Check out this new version from our repositories. API Javadoc is already available. Your suggestions will be very precious for us, so download and try KeLP 2.2.2!

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

Release 0.7.0

• Update to TensorFlow r1.4.

• Tutorials in the documentation on:

• Interfacing with Keras

• Saving/loading models

• How to build a variety of regressors with Aboleth

• New prediction module with some convenience functions, including freezing the weight samples during prediction.

• Bayesian convolutional layers with accompanying demo.

• Allow the number of samples drawn from a model to be varied by using placeholders.

• Generalise the feature embedding layers to work on matrix inputs (instead of just column vectors).

• Numerous numerical and usability fixes.

About: DynaML is a Scala environment for conducting research and education in Machine Learning. DynaML comes packaged with a powerful library of classes implementing predictive models and a Scala REPL where one can not only build custom models but also play around with data work-flows.

Initial Announcement on mloss.org.

About: A library of scalable Bayesian generalised linear models with fancy features

• 1.0 release!
• Now there is a random search phase before optimization of all hyperparameters in the regression algorithms. This improves the performance of revrand since local optima are more easily avoided with this improved initialisation
• Regression regularizers (weight variances) associated with each basis object, this approximates GP kernel addition more closely
• Random state can be set for all random objects
• Numerous small improvements to make revrand production ready
• Final report
• Documentation improvements

About: A Java Toolbox for Scalable Probabilistic Machine Learning.

• Added sparklink module implementing the integration with Apache Spark. More information here.
• Fluent pattern in latent-variable-models
• Predefined model implementing the concept drift detection

Detailed information can be found in the toolbox's web page

About: A Python based library for running experiments with Deep Learning and Ensembles on GPUs.

Initial Announcement on mloss.org.

About: Software for Automatic Construction and Inference of DBNs Based on Mathematical Models

Initial Announcement on mloss.org.

About: Efficient and Flexible Distributed/Mobile Deep Learning Framework, for python, R, Julia and more

This version comes with Distributed and Mobile Examples

About: Apache Mahout is an Apache Software Foundation project with the goal of creating both a community of users and a scalable, Java-based framework consisting of many machine learning algorithm [...]

Apache Mahout introduces a new math environment we call Samsara, for its theme of universal renewal. It reflects a fundamental rethinking of how scalable machine learning algorithms are built and customized. Mahout-Samsara is here to help people create their own math while providing some off-the-shelf algorithm implementations. At its core are general linear algebra and statistical operations along with the data structures to support them. You can use is as a library or customize it in Scala with Mahout-specific extensions that look something like R. Mahout-Samsara comes with an interactive shell that runs distributed operations on a Spark cluster. This make prototyping or task submission much easier and allows users to customize algorithms with a whole new degree of freedom. Mahout Algorithms include many new implementations built for speed on Mahout-Samsara. They run on Spark 1.3+ and some on H2O, which means as much as a 10x speed increase. You’ll find robust matrix decomposition algorithms as well as a Naive Bayes classifier and collaborative filtering. The new spark-itemsimilarity enables the next generation of cooccurrence recommenders that can use entire user click streams and context in making recommendations.

About: xgboost: eXtreme Gradient Boosting It is an efficient and scalable implementation of gradient boosting framework. The package includes efficient linear model solver and tree learning algorithm. The package can automatically do parallel computation with OpenMP, and it can be more than 10 times faster than existing gradient boosting packages such as gbm or sklearn.GBM . It supports various objective functions, including regression, classification and ranking. The package is made to be extensible, so that user are also allowed to define there own objectives easily. The newest version of xgboost now supports distributed learning on various platforms such as hadoop, mpi and scales to even larger problems

• Distributed version of xgboost that runs on YARN, scales to billions of examples

• Direct save/load data and model from/to S3 and HDFS

• Feature importance visualization in R module, by Michael Benesty

• Predict leaf index

• Poisson regression for counts data

• Early stopping option in training

• Native save load support in R and python

• xgboost models now can be saved using save/load in R

• xgboost python model is now pickable

• sklearn wrapper is supported in python module

• Experimental External memory version

About: streamDM is a new open source data mining and machine learning library, designed on top of Spark Streaming, an extension of the core Spark API that enables scalable, high-throughput, fault-tolerant stream processing of data streams.

Initial Announcement on mloss.org.

About: Hivemall is a scalable machine learning library running on Hive/Hadoop.

• Supported Matrix Factorization
• Added a support for TF-IDF computation
• Supported AdaGrad/AdaDelta
• Supported AdaGradRDA classification
• Added normalization scheme

About: rabit (Reliable Allreduce and Broadcast Interface) is a light weight library that provides a fault tolerant interface of Allreduce and Broadcast for portable , scalable and reliable distributed machine learning programs. Rabit programs can run on various platforms such as Hadoop, MPI and no installation is needed. Rabit now support kmeans clustering, and distributed xgboost: an extremely efficient disrtibuted boosted tree(GBDT) toolkit.

Initial Announcement on mloss.org.

About: Scalable tensor factorization

• Improve (speed up) initialization of A by summation

About: Efficient implementation of Semi-Stochastic Gradient Descent algorithm (S2GD) for training logistic regression (L2-regularized).

Initial Announcement on mloss.org.

About: Encog is a Machine Learning framework for Java, C#, Javascript and C/C++ that supports SVM's, Genetic Programming, Bayesian Networks, Hidden Markov Models and other algorithms.

Changes for Encog 3.2:

Issue #53: Fix Out Of Range Bug In BasicMLSequenceSet. Issue #52: Unhandled exception in Encog.Util.File.ResourceLoader.CreateStream (ResourceLoader.cs) Issue #50: Concurrency bugs in PruneIncremental Issue #48: Unit Tests Failing - TestHessian Issue #46: Couple of small fixes - Temporal DataSet and SCG training Issue #45: Fixed EndMinutesStrategy to correctly evaluate ShouldStop after the specified number of minutes have elapsed. Issue #44: Encog.ML.Data.Basic.BasicMLDataPairCentroid.Add() & .Remove() Issue #43: Unit Tests Failing - Matrix not full rank Issue #42: Nuget - NuSpec Issue #36: Load Examples easier

About: RLLib is a lightweight C++ template library that implements incremental, standard, and gradient temporal-difference learning algorithms in Reinforcement Learning. It is an optimized library for robotic applications and embedded devices that operates under fast duty cycles (e.g., < 30 ms). RLLib has been tested and evaluated on RoboCup 3D soccer simulation agents, physical NAO V4 humanoid robots, and Tiva C series launchpad microcontrollers to predict, control, learn behaviors, and represent learnable knowledge. The implementation of the RLLib library is inspired by the RLPark API, which is a library of temporal-difference learning algorithms written in Java.

Current release version is v2.0.