EARL London 2019

Deep learning with neural networks has changed the way many machine learning problems are now solved, but until recently was difficult to use outside of large research groups. The Keras deep learning library has dramatically reduced the barrier to entry by providing a simple interface to lower level libraries, such as TensorFlow or Theano. This hands-on workshop will guide you through creating neural networks with the R interface to Keras. We will cover some theoretical background and discuss the types of problem for which neural networks are appropriate. The focus is on implementation, and you will build your own networks in guided exercises.

Attendees should be familiar with the R language, and comfortable manipulating R data structures, such as matrices and arrays.

The Package Development in R workshop is aimed at R programmers who are looking to add rigour to their development by building R packages in a controlled, scalable and commercially viable manner. Attendees will learn the state of the art for designing, building and maintaining R packages. The workshop will maintain a practical and current focus using popular R packages such as devtools, roxygen2 and testthat.

• Participants should have a good working knowledge of the R language but experience of package building will not be assumed • Participants should have a recent version of R installed and, if working on a Windows Desktop they must have an appropriate version of Rtools installed. Registrants will be provided with a list of R packages to install prior to the workshop.