Scalable Probabilistic Programming for Learning and Acting in Complex Domains

Abstract

The research objective of this seedling effort was to create probabilistic programming languages (PPLs) that are scalable and expressive enough to serve as the foundation for complex agents (such as autonomous robots) that must blend low-level perception and high-level reasoning. The technical strategy was to improve the theory and implementation of PPLs through a mix of language refinement, novel inference algorithm development, and hardware acceleration. Key results from this project include, strong results on simulator development, development of the Runner-Chaser model, development of probabilistic Schelling games, and failed experiments on high-performance, low-level perception.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 2019
Accession Number
AD1074255

Entities

People

  • David Wingate

Organizations

  • Brigham Young University

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Algorithms
  • Artificial Intelligence
  • Artificial Neural Networks
  • Bayesian Networks
  • Cognition
  • Cognitive Science
  • Computational Science
  • Computer Programming
  • Detection
  • Information Processing
  • Information Science
  • Information Systems
  • Machine Learning
  • Monte Carlo Method
  • Neural Networks
  • Probabilistic Models
  • Probability
  • Probability Distributions
  • Programming Languages
  • Psychological Theory
  • Random Variables
  • Reasoning
  • Simulations
  • Simulators
  • Unmanned Aerial Vehicles

Fields of Study

  • Computer science

Readers

  • Computational Linguistics
  • Distributed Systems and Data Platform Development

Technology Areas

  • AI & ML
  • Autonomy
  • Autonomy - Autonomous System Control