COGNITIVELY-INSPIRED AGILE INFORMATION AND KNOWLEDGE MODELLING (CALM)

Abstract

The impetus for this project stems from a quest to create the formal structure of an information and knowledge modelling approach for measuring the security risk posed by embedded electronic systems in sensitive industries. The approach needs to combine innovative and agile data acquisition, signal processing, feature extraction, data fusion, and deep neural networks for data classification, segmentation, interpretation and decision making. Hybrid and Deep learning mechanisms such as deep neural networks are being successfully deployed in a wide range of applications, but learning methodologies are still heavily reliant on human intervention throughout initial settings, design and training processes (application life cycle). The solutions normally become training dataset and application specific. Any changes to the data specifications, data collection conditions and signals or application re-orientation usually lead to re-design of the entire Neural Network Architecture (NNA). The re-design of the NNA leads to major time, cost and computational inefficiencies. The sensitive electronic/microelectronic quality and security assurance industry needs a portable and flexible solution capable of handling subjects of interest which are normally of high variety and throughput. The speed and accuracy of recognition is therefore vital. We wish to explore the creation of Agile Capsule Neural Network Systems to overcome existing shortcomings of deep learning models (see Figure 1).

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 12, 2021
Source ID
FA86552017051

Entities

People

  • Alireza Mousavi

Organizations

  • Air Force Office of Scientific Research
  • Brunel University London
  • United States Air Force

Tags

Fields of Study

  • Computer science

Readers

  • Distributed Systems and Data Platform Development
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • AI & ML - Neural Networks
  • Microelectronics