University of Michigan - Ann Arbor

Deep Reinforcement Learning (DRL) is a technology that combines the two, creating a sequential reinforcement learning process, in which deep learning determines the action taken at every stage.

Deep Reinforcement Learning 

An Introduction to Deep Reinforcement Learning
https://arxiv.org/pdf/1811.12560.pdf

Reference for Deep Reinforcement Learning

There are different types of deep reinforcement learning methods.

If we use either value functions or policies to act on the environment, the method is called Model-free reinforcement learning:
   * Value-based methods
   * Policy-based methods

If we make use of models of the environment, it is referred to as Model-based reinforcement learning.

Deep Reinforcement Learning Methods

https://missinglink.ai/guides/neural-network-concepts/complete-guide-deep-reinforcement-learning-concepts-process-real-world-applications/

Complete Guide to Deep Reinforcement Learning: Concepts, Process, and Real World Applications

http://icaps18.icaps-conference.org/fileadmin/alg/conferences/icaps18/summerschool/lectures/Lecture5-rl-intro.pdf

Slides from CMU: Introduction to Deep Reinforcement Learning

Deep Reinforcement Learning (DRL) is data inefficient and may require large amounts of data through millions of iterations to give results for simple tasks. To circumvent this, simulated environments are often used to train DRL models. There are some major practical gaps between simulated and real environments and this makes the models difficult to operate feasibly in the real-world. This is why DRL is used widely in closed environments like video games, but is difficult to apply in real-world environments. Some organizations use this by opting for a deep learning platform (platforms with large data sets and documentation easily available) to help them implement DRL.

Using DRL in Real-World Environments

1. Accelerating Human Learning with Deep Reinforcement Learning
2. Using deep reinforcement learning for personalizing review sessions
on e-learning platforms with spaced repetition
3. A Spaced, Interleaved Retrieval Practice Tool that is Motivating and Effective
4. DAS3H: Modeling Student Learning and Forgetting for Optimally Scheduling Distributed Practice of Skills
5. Deep-IRT: Make Deep Learning Based Knowledge Tracing
Explainable Using Item Response Theory
6. Knowledge Query Network for Knowledge Tracing
7. Knowledge Tracing Machines: Factorization Machines for Knowledge Tracing
8. A Self-Attentive model for Knowledge Tracing
9. Context-Aware Attentive Knowledge Tracing
10. Augmenting Knowledge Tracing by Considering Forgetting
Behavior

Learn Before

Related