Model Explanation by Optimal Selection of Teaching Examples

Abstract

Our contributions to XAI can be summarized by four lines of work: (1) developing novel XAI techniques, (2) demonstrating successful XAI systems with empirical validation, (3) advancing our understanding of the psychology of explainability, and (4) contributing to the mathematical foundations of XAI. The theme of our approach is to incorporate human inference into all aspects of XAI.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2022
Accession Number
AD1186460

Entities

People

  • Patrick Shafto
  • Scott Cheng-hsin Yang

Organizations

  • Rutgers University

Tags

Communities of Interest

  • Autonomy
  • Human Systems

DTIC Thesaurus Topics

  • Abstracts
  • Accuracy
  • Air Force
  • Air Force Research Laboratories
  • Algorithms
  • Artificial Intelligence
  • Behavioral Sciences
  • Cognitive Science
  • Data Science
  • Deep Learning
  • Experimental Design
  • Formal Languages
  • Gaussian Processes
  • Information Processing
  • Information Science
  • Information Systems
  • Machine Learning
  • Neural Networks
  • Psychological Theory
  • Psychology
  • Reliability
  • Statistical Analysis
  • Textbooks

Fields of Study

  • Computer science

Readers

  • Artificial Intelligence
  • Theoretical Analysis.

Technology Areas

  • AI & ML