A Series of Unlikely Events: Learning from Sequential Behavior for Activity Based Intelligence and Modeling Human Expertise

Abstract

Modeling patterns of sequential behavior is a task that underlies numerous difficult artificial intelligence tasks: How do I detect when adversaries are deviating from normal routines? How can I predict where a ship is going to dock? How can I automate the teaching of novice analysts to perform complex tasks as if they were experts? In this work, we use a class of techniques called Imitation Learning (IL) to model sequential behavior to answer questions like these and others.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2020
Accession Number
AD1110464

Entities

People

  • Dan Decapria
  • Eric Heim
  • Jake Oaks
  • Jay Palat
  • Jonathan Hoyle

Organizations

  • Carnegie Mellon University

Tags

Communities of Interest

  • Ground and Sea Platforms

DTIC Thesaurus Topics

  • Applied Computer Science
  • Artificial Intelligence
  • Computer Science
  • Demonstrations
  • Education
  • Information Processing
  • Information Systems
  • Learning
  • Nautical
  • Observation
  • Psychology
  • Reinforcement Learning
  • Schools
  • Scientists
  • Theoretical Computer Science

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Artificial Intelligence
  • Regression Analysis.

Technology Areas

  • AI & ML