Symbolic Reasoning meets Deep Learning: Toward Building Machines that Think

Abstract

This project is titled "Symbolic Reasoning meets Deep Learning: Toward Building Machines that Think". The goal of the project is to develop methods toward a general and versatile symbolic reasoning engine capable of supporting a broad range of Artificial Intelligence tasks and applications. Symbolic reasoning means performing inference through the manipulation of symbols as epitomized by mathematical reasoning. The main approach of this project is a synthesis between symbolic reasoning, which is discrete in nature, and deep learning, which excels at continuous vector representations. This projectconsists of two thrusts: neural heuristic search, which seeks to integrate deep learning with symbol manipulation, and 4D symbol grounding, which seeks to learn the bi-directional links between semantic structures and dynamic visual scenes. If successful, this project will majorly advance automated understanding of visual and textual data. The Principal Investigator is Dr. Jia Deng at Princeton University, and the total funds requested is $509,298.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 31, 2020
Source ID
N000142012634

Entities

People

  • Jia Deng

Organizations

  • Office of Naval Research
  • Trustees of Princeton University
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Neural Network Machine Learning.
  • Technical Research and Report Writing.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks