Workshop host: Leon Eyrich Jessen
ANNs form the basic unit of deep learning and are a powerful tool in predictive modeling, albeit not without pitfalls. In this workshop, you will…
- get an introduction to Artificial Neural Networks (ANNs) in R with Keras and TensorFlow
- be working with conceptually understanding the inner workings of ANNs
- touch upon the main pitfall of ANNs, model over-fitting
The workshop assumes basic R/Data Science skills as well as knowledge of rmarkdown and the RStudio IDE. Please note, the workshop is very hands-on oriented, so expect to get your fingers dirty!
Running 00_setup_workshop_packages.R will setup the needed
environment.
Total time 4h:
-
00.00 - 00.20 (20min) Talk: Introduction to Artificial Neural Networks
-
00.20 - 00.50 (30min) Exercise 1: Prototyping an ANN in R
-
00.50 - 01.00 (10min) Break I
-
01.00 - 01.05 (5min) Exercise 1 walk-through
-
01.05 - 01.10 (5min) Brief talk: Introduction to TensorFlow/Keras in R part 1
-
01.10 - 01.30 (20min) Exercise 2: TensorFlow Playground
-
01.30 - 01.40 (10min) Exercise 2 walk-through
-
01.40 - 01.45 (5min) Brief talk: Introduction to TensorFlow/Keras in R part 2
-
01.45 - 01.50 (5min) Brief talk: Session 1 Summary and Q&A
-
01.50 - 02.00 (10min) Break II
-
02.00 - 02.30 (30min) Exercise 3: Hello Keras
-
02.30 - 02.45 (15min) Brief talk and exercise 3 walk-through: A bit more on Keras
-
02.45 - 03.00 (15min) Break III
-
03.00 - 03.30 (30min) Exercise 4: Now, you must choose between:
-
03.30 - 03.55 (25min) Exercise 5: A Case Story
-
03.55 - 04.00 (5min) Brief talk and exercise 5 walk-through: Session 2 Summary