Author: Wouter van Heeswijk, PhD
-
![Instead of benchmark optimization- and machine learning algorithms against each other, we should consider how they can strengthen each other [Photo by Wedding Dreamz on Unsplash]](https://towardsdatascience.com/wp-content/uploads/2023/12/0zj0ikUUckUyLtVwF-scaled.jpg)
How pattern detection and pattern exploitation might elevate each other to a new level
15 min read -
Learn more reliable, robust, and transferable policies by adding entropy bonuses to your algorithm
10 min read -
![And...action! [Photo by Jakob Owens on Unsplash]](https://towardsdatascience.com/wp-content/uploads/2023/08/0Mkxm4Grw4SHsV6O1-scaled.jpg)
Action spaces, particularly in combinatorial optimization problems, may grow unwieldy in size. This article discusses…
18 min read -
A gradient-based reinforcement learning algorithm to learn deterministic policies for continuous action spaces
12 min read -
A Python implementation of Q-learning to solve the Taxi-v3 environment from OpenAI Gym in an…
8 min read -
![Rock-paper-scissors would be a dull affair with deterministic policies [Photo by Marcus Wallis on Unsplash]](https://towardsdatascience.com/wp-content/uploads/2023/02/0VA5suMpH7bXhNQvq-scaled.jpg)
Why we let randomness dictate our action selection in Reinforcement Learning
7 min read -
Harness yourself against these shortcomings encountered in everyday RL algorithms
9 min read -
![Empirical performance, averaged over 57 Atari games. Rainbow DQN strongly outperforms individual DQN techniques, both in terms of rate of learning and final performance [source: DeepMind paper]](https://towardsdatascience.com/wp-content/uploads/2022/12/1Z9K2oKzLpXwse8-QB-y8Yw.png)
What happens if the most successful techniques in Deep Q-Learning are combined into a single…
14 min read -
The journey from REINFORCE to the go-to algorithm in continuous control
16 min read