RcppEnsmallen v0.2.13.0.1 Released - Early Stopping with Lambdas and Update Deployment Automation

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. After a three month hiatus, the Ensmallen Development Team released a new patch that enabled callbacks for early stopping procedures to receive a lambda function. Outside of that, the group worked on improving automation on new update releases.

Read More…

uiucthemes v0.3.1 Released - Upstream compatibility fixes and maintenance

The uiucthemes package is an organizational R package that provides unified brand templates based on the University of Illinois at Urbana-Champaign’s (UIUC) identity guidelines. In particular, the package houses different Beamer and HTML slide templates hooked into R Markdown. As time spins onward, uiucthemes needed a compatibility fix to work with the latest version of rmarkdown. In particular, newer versions of rmarkdown threw the following warning: “citation_package = ’none’ was deprecated; please use ‘default’ instead.

Read More…

RcppEnsmallen v0.2.12.1.1 Released - QoL and compatibility improvements

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. For the third update of 2020, the Ensmallen Development Team focused on improving the quality of the package and fixing issues with Armadillo compatibility version compatibility. Details of all changes can be found in the list below.

Read More…

edina v0.1.1 Released: Estimate an Exploratory Deterministic Input, Noisy 'and' Gate Model!

The edina package provides a high-performance implementation of the Exploratory Deterministic Input, Noisy ‘and’ Gate Model (EDINA) described by Chen et al. (2018). Exploratory models estimate the Q-matrix alongside other parameters within a cognitive diagnostic model (CDM) unlike confirmatory models, which require a pre-specified expert Q matrix. For more information, please see the full NEWS release below. edina news file entry for version 0.1.1 (2020-03-25) Features Provides a high-performing modeling routine for the Exploratory Deterministic Input, Noisy “And” Gate model.

Read More…

cIRT v1.3.1 Released: More Modernization and Maintenance Release

The cIRT package (CRAN, GitHub) is home to the implementation of Choice Item Response Theory (cIRT) described by Culpepper and Balamuta (2015). cIRT jointly models the accuracy of cognitive responses and item choices within a bayesian hierarchical framework. After over a year, we’re back for another maintenance release! I suppose this goes to show that science requires maintenance. In particular, this release addressed a CRAN issue with respect to how data is stored in the package.

Read More…

errum v0.0.3 Released: Model an Exploratory Reduced Reparameterized Unified Model!

The errum package provides a high-performance implementation of the Exploratory reduced Reparameterized Unified Model (ErRUM) described by Culpepper and Chen (2018). Development of this model followed from Chen et al. (2018)’s work on estimating the Q-matrix within a cognitive diagnostic model (CDM). With the package now on CRAN, this is the first time an exploratory model has been made available publically! For more information, please see the full NEWS release below.

Read More…

errorist v0.1.0 Released: Improved automatic warning searches

errorist: one who holds to and propagates error — Merriam Webster The goal of errorist is to automatically search errors and warnings the second they arise. To accomplish this, errorist automatically applies a set of environment hooks on package load that retrieve errors and warnings, which occur during the execution of code, so that they can be automatically searched for solutions. For the fourth ever errorist update, support was significantly improved for handling multiple warning message searches.

Read More…

RcppEnsmallen v0.2.11.3.1 Released - Memory leak fix and other goodies!

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 start of 2020 is off to a bang with the Ensmallen Development Team releasing a new update! In this update, many bugs were squashed; though, perhaps more importantly: memory leaks with SGD were squashed!

Read More…

searcher v0.0.5 Released - Search Twitter and Search Experience Customizations!

The searcher package (CRAN, GitHub) provides a search interface to look up terms on popular websites such as Google, Bing, DuckDuckGo, Startpage, Twitter, StackOverflow, RStudio Community, GitHub, and BitBucket. Upon calling a search_*() function, a browser window will open with the search results for the query. That said, a new year means new searcher capabilities! The searcher package received its biggest ever overhaul to date in its quest to provide a stable and customizable API.

Read More…

RcppEnsmallen v0.2.11.1.1 Released - New Optimizers and Documentation Improvements!

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. With Comprehensive R Archival Network on vacation for two weeks, we missed out on the Ensmallen Development Team Christmas release until now. The latest build adds three new optimizers: AdaBound, AMSBound, and Lookahead.

Read More…