Research Methods

Distilling Information from a Flood: A Possibility for the Use of Meta-Analysis and Systematic Review in Machine Learning Research

The current flood of information in all areas of machine learning research, from computer vision to reinforcement learning, has made it difficult to make aggregate scientific inferences. It can be challenging to distill a myriad of similar papers …

Ethical Challenges in Data-Driven Dialogue Systems

The use of dialogue systems as a medium for human-machine interaction is an increasingly prevalent paradigm. A growing number of dialogue systems use conversation strategies that are learned from large datasets. There are well documented instances …

Benchmarking and Evaluation in Inverse Reinforcement Learning

Benchmarks are particularly useful for characterizing algorithms and determining their usefulness in different settings. Here I highlight the need for more standardization of performance evaluations in inverse reinforcement learning.

Reproducibility and Replicability in Deep Reinforcement Learning (and Other Deep Learning Methods)

In recent years, significant progress has been made in solving challenging problems using deep learning. Reproducing existing work and accurately judging the improvements offered by novel methods is vital to sustaining this progress. Unfortunately, …

Show Me the Data! On the Reproducibility of Policy Gradient Methods for Continuous Control

A brief discussion on some difficulties that students may encounter in reproducing modern policy gradient methods in continuous control tasks and best practices for writing papers on these methods.

Practical Tutorial on Policy Gradients for Continuous Control

A practical tutorial session on implementing and running policy gradient methods for continuous control tasks.