COSMIC: Commonsense Machine Intelligence

Abstract

The scope of this effort is to develop a Machine Common Sense (MCS) learning framework for Obvious Plans and Inferences for Common Sense (OPICS) service for Artificial Intelligence (AI) and robotic systems. This MCS learning framework will learn MCS planning and inference capabilities about objects, agents, and places equivalent to the capabilities of an 18-month-old infant. The proposed research addresses five fundamental challenges toward MCS: (1) lack of comprehensive KB for causal and multimodal commonsense knowledge, (2) lack of commonsense reasoning engines, (3) lack of universal commonsense embeddings, (4) MCS dataset bottlenecks, and (5) the major discrepancy between existing ML paradigms and MCS-driven learning of humans

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 13, 2022
Source ID
N660011924031

Entities

People

  • Yejin Choi

Organizations

  • Defense Advanced Research Projects Agency
  • Naval Information Warfare Center Pacific
  • University of Washington

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • Autonomy