Foundations of Sequential Learning

Abstract

This report summarizes the research done under FA8750-16-2-0173. This research advanced understanding of bandit algorithms and exploration in Markov Decision Processes (MDPs). New algorithms and theory were proposed for bandits with periodic payoff multipliers and arms with costs. Exploration and transfer learning algorithms were evaluated for MDPs.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 2018
Accession Number
AD1047509

Entities

People

  • Cynthia Rudin
  • Kamesh Munagala
  • Ronald Parr

Organizations

  • Duke University

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Bayesian Networks
  • Contracts
  • Government Procurement
  • Governments
  • Information Exchange
  • Law
  • Learning
  • Probability
  • Reinforcement Learning
  • Sampling
  • Security
  • Standards
  • Technical Information Centers
  • Transfer Functions
  • Universities

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Distributed Systems and Data Platform Development
  • Operations Research