Can Discovering Boost Learning? Improving the Quality of a Machine Learning Model through Discovering Hidden Structure Among Data

Abstract

The PI has successfully completed the research proposed. In this project the PIs studied how the performance of a machine learning model can be boosted through incorporating a discovery engine aiming at finding hidden/missing structure or information. The team used an unsupervised node-ranking model that considers not only the attributes of nodes in a graph but also the incompleteness of the graph structure. They showed that by discovering the structure of networks they were able to perform better node ranking. They then used deep neural network (DNN)( based solution as a ranking model. The rich representation capability of the DNN structure together with a novel design of the discover objectives allow the proposed model to significantly outperform the state-of-the-art ranking solutions. There were 3 peer reviewed publications as a direct result of this grant award.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 02, 2019
Accession Number
AD1077389

Entities

People

  • Shou-De Lin

Organizations

  • National Taiwan University

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Artificial Intelligence Software
  • Artificial Neural Networks
  • Data Mining
  • Deep Learning
  • Department Of Defense
  • Electronic Mail
  • Information Science
  • Learning
  • Machine Learning
  • Maximum Likelihood Estimation
  • Neural Networks
  • New York
  • Optimization
  • Probability
  • Probability Distributions
  • Random Walk
  • Semi-Supervised Learning
  • Social Networks
  • Supervised Machine Learning
  • Training
  • Unsupervised Machine Learning

Fields of Study

  • Computer science

Readers

  • Neural Network Machine Learning.
  • Research Science/Academic Research

Technology Areas

  • AI & ML
  • AI & ML - Information Retrieval
  • AI & ML - Neural Networks