RcppEnsmallen v0.2.18.0.1 Released - Gradient Stability Techniques and Bugfixes!

The RcppEnsmallen package brings to R a Header-Only C++ Mathematical Optimization Library for Armadillo. In particular, Ensmallen is a templated C++ mathematical optimization library (by the MLPACK team) that provides a simple set of abstractions for writing an objective function to optimize. The fall update of the ensmallen library has been released by the Ensmallen Development Team. Within this release, we see improvements to gradient stability with the inclusion of a gradient clipping and norm scaling callback based on “On the difficulty of training Recurrent Neural Networks” by Razvan Pascanu, Tomas Mikolov, Yoshua Bengio.

Read More…

sitmo v2.0.2 Released - Bag of Bugfixes and Modernizations

The objective of sitmo (CRAN, GitHub) is to provide the means for performing parallel draws from a Parallel Pseudo Random Number Generator (PPRNG). To that end, the package is largely a bridge into the PPRNGs developed by the sitmo consultancy agency. After two years without a new release, sitmo has been updated to address an issue highlighted by Yihui in knitr issue #2057 regarding unbalanced code chunk delimiters within the big_crush_test.

Read More…

edmdata v1.2.0 Released: New Oracle Strategy Sets and Probability Assessment Data

The edmdata package (GitHub, Website) provides data sets from various administered assessments that can be used in diagnostic modeling. These data sets have been analyzed in various papers that introduced new methodology as part of the application section. With the release of Chen, Yinghan, Liu, Y., Culpepper, S. A., & Chen, Y. (2021), we have incorporated into edmdata the Elementary Probability Theory Assessment from the {pks} package. When compared to {pks}, the data has been re-formatted from the single probability data frame object into a single trial matrix containing only part one of the assessment.

Read More…

searcher v0.0.6 Released - Search Ecosia and Rseek!

The searcher package (CRAN, GitHub) provides a search interface to look up terms on popular websites such as Google, Bing, DuckDuckGo, Startpage, Ecosia, Rseek, Twitter, StackOverflow, RStudio Community, GitHub, and BitBucket. Upon calling a search_*() function, a browser window will open with the search results for the query. After a year of dormacy, the searcher package is back with a new update! In this update, we’ve added two requested searcher portals: Ecosia and Rseek.

Read More…

RcppEnsmallen v0.2.17.0.1 Released - Support for Multiobjective Optimizers!

The RcppEnsmallen package brings to R a Header-Only C++ Mathematical Optimization Library for Armadillo. In particular, Ensmallen is a templated C++ mathematical optimization library (by the MLPACK team) that provides a simple set of abstractions for writing an objective function to optimize. Coming in right after the Fourth of July, we have the second minor release from Ensmallen Development Team. This release contains work done by Nanubala Gnana Sai, who worked extensively on a framework for Multiobjective Optimizers during Google Summer of Code 2021.

Read More…

RcppEnsmallen v0.2.16.1.1 Released - Better Unit Tests and Bug Fixes!

The RcppEnsmallen package brings to R a Header-Only C++ Mathematical Optimization Library for Armadillo. In particular, Ensmallen is a templated C++ mathematical optimization library (by the MLPACK team) that provides a simple set of abstractions for writing an objective function to optimize. The Ensmallen Development Team second release of 2021 has been packaged and sent to CRAN. The main changes with this version are more or less behind the scenes with improvements to unit testing (not activated in the CRAN release) and a few bug fixes.

Read More…

edmdata v1.1.0 Released: New Low Attribute Oracle Q and Documentation Fixes

The edmdata package (GitHub, Website) provides data sets from various administered assessments that can be used in diagnostic modeling. These data sets have been analyzed in various papers that introduced new methodology as part of the application section. In this release, we add one additional oracle Q matrix and fix an issue with a prior oracle Q matrix that was included in the package. Moreover, we addressed CRAN remarks regarding DOI links.

Read More…

RcppEnsmallen v0.2.15.0.1 Released - New Callback: Optimization Summary Report

The RcppEnsmallen package brings to R a Header-Only C++ Mathematical Optimization Library for Armadillo. In particular, Ensmallen is a templated C++ mathematical optimization library (by the MLPACK team) that provides a simple set of abstractions for writing an objective function to optimize. On election day in the United States, the Ensmallen Development Team made available a new version of Ensmallen with a killer new feature: Optimization Summary Reports. From the documentation, we have:

Read More…

msos v1.2.0 Released: Unbalanced data bug fix and a pkgdown site!

The msos R package serves as the companion package to Professor Emeritus John Marden’s Multivariate Statistics: Old School, which is used in STAT 571: Multivariate Analysis course at the University of Illinois at Urbana-Champaign (UIUC). In this hotfix, a long standing bug in the lda() routine that affected unbalanced data was corrected thanks to Trevor Park who reported the issue! Moreover, the package has switched from being tested with TravisCI to using GitHub Actions and now has its own pkgdown website at: https://coatless.

Read More…

RcppEnsmallen v0.2.14.2.1 Released - New NSGA2 optimizer and Bug Fixes!

The RcppEnsmallen package brings to R a Header-Only C++ Mathematical Optimization Library for Armadillo. In particular, Ensmallen is a templated C++ mathematical optimization library (by the MLPACK team) that provides a simple set of abstractions for writing an objective function to optimize. There are a lot of goodies in this release. Portions of the goodies are thanks to the excellent job done by the Ensmallen Development Team. For this update, they’ve add the NSGA2 optimizer and addressed a bunch of edge case bugs.

Read More…

All Posts by Category or Tags.