Home › Categories › Technology› Instance-Based Generalization In Reinforcement Learning

Instance-Based Generalization In Reinforcement Learning.pdf

2011.01089.pdf
Preview of Instance-Based Generalization in Reinforcement Learning
🔗 Source: arxiv.org
📊 Size: 780 KB
📄 Pages: 21 pages
⬇️ Downloads: 147

Summary

Agents trained via deep reinforcement learning (RL) fail to generalize to unseen environments despite sharing the same underlying dynamics as the training levels. This is due to the agent learning instance-specific policies instead of generalizable ones, resulting from maximizing expected rewards and inducing undesired speed-running policies. The effective Markov dynamics the agent observes during training are significantly changed when reusing instances, leading to sub-optimal policies on the training set. Generalization bounds to the value gap in train and test environments are provided based on the number of training instances. A proposed solution involves training a shared belief representation over an ensemble of specialized policies and computing a consensus policy for data collection, disallowing instance-specific exploitation. Experimental validation is provided using the CoinRun benchmark.

Download File

Description

Agents trained via deep reinforcement learning (RL) fail to generalize to unseen environments despite sharing the same underlying dynamics as the training...

Technical Information

  • File Format: PDF
  • File Size: 780 KB
  • Pages: 21
  • Language: EN
  • Total Downloads: 147
  • Last Updated: 3 days ago

Document Overview

This PDF document about Instance-Based Generalization in Reinforcement Learning provides comprehensive information and guidance. Whether you're a beginner or advanced user, this resource offers valuable insights into Instance-Based Generalization in Reinforcement Learning.

Related Topics

If you're interested in Instance-Based Generalization in Reinforcement Learning, you might also want to explore:

Instance-Based guideInstance-Based tutorialGeneralization guideGeneralization tutorialReinforcement guideReinforcement tutorialLearning guideLearning tutorial

Related Categories

Business & Finance
41204 documents
Education & Learning
22711 documents
Health & Wellness
8956 documents

Related Documents

Preview of Reinforcement Learning
PDF file

Reinforcement Learning

1.42 MBGabriel Turinici, P. BrugiereEN23 pages
Reinforcement learning algorithms have been successfully used to find good strategies based on available information, but few works have...
Download
Preview of Reinforcement Learning
PDF file

Reinforcement Learning

1.37 MBEN73 pages
Reinforcement learning offers a framework for designing sophisticated robot behaviors through trial-and-error interactions with the environment,...
Download
Preview of Networks of Relations: Distributed Learning and Generalization
PDF file

Networks of Relations: Distributed Learning and Generalization

114 KBEN6 pages
Networks of Relations for Representation, Learning, and Generalization propose knowledge representation as a network of relations between variables,...
Download
Preview of Reinforcement Learning for Explainable Recommendation
PDF file

Reinforcement Learning for Explainable Recommendation

1.38 MBYikun Xian, Zuohui Fu, S. Muthukrishnan, Gerard de Melo, and Yongfeng ZhangEN10 pages
A method for explainable recommendations using knowledge graph reasoning, providing interpretable paths.
Download
Preview of Mitigating Catastrophic Forgetting in Reinforcement Learning
PDF file

Mitigating Catastrophic Forgetting in Reinforcement Learning

2.49 MBChristos Kaplanis, Murray Shanahan, Claudia ClopathEN14 pages
Unlike humans, artificial neural networks suffer from catastrophic forgetting, where new learning erases previously acquired knowledge. We show that...
Download

Download Instance-Based Generalization in Reinforcement Learning eBooks for free and learn more about Instance-Based Generalization in Reinforcement Learning. These books contain exercises and tutorials to improve your practical skills, at all levels!

Not satisfied with this document? We have related documents to Instance-Based Generalization in Reinforcement Learning, try searching with similar keywords: Instance-Based Generalization in Reinforcement Learning, Instance Segmentation via Instance Categories: A Simple and Flexible Framework, learning and generalization, Networks of Relations: Distributed Learning and Generalization, Solutions Reinforcement And Answers Reinforcement, POET: Open-ended Reinforcement Learning through Unbounded Invention of Learning Challenges and their Solutions, Poet Open Ended Reinforcement Learning Through Unbounded Invention Of Learning Challenges And Their Solutions, A Sample Html Instance

You can download PDF versions of the user's guide, manuals and ebooks about Instance-Based Generalization in Reinforcement Learning, you can also find and download for free A free online manual (notices) with beginner and intermediate, Downloads Documentation, You can download PDF files (or DOC and PPT) about Instance-Based Generalization in Reinforcement Learning for free, but please respect copyrighted ebooks.