San Diego Section - The Institute of Electrical and Electronics Engineers, Inc.

Home
Section Information
Societies
Announcements
Upcoming Events
Conferences
Technology Time Machine
Download Presentations
Consultants Network
Student Branches
Chair's Message
Jobs
Contact


Officer Login
Email Address:
Password:

Don't know your password?

Feature Object Extraction - IEEE AESS Distinguished Lecture

on June 24, 2020 from 6:00 PM to 7:30 PM

Speaker Kathleen Kramer and Stephen Stubberud
Topic Info

Information assurance, also referred to as cyber security, is the process of protecting information from theft, destruction, or manipulation. Cyber threats can be either from internal or external sources, sudden or taking time to develop, such as a slow denial of service (DOS) attack. Some techniques have been developed to behave as sensors to quickly assess elements of attacks that rely on a decision engine to fuse the information to estimate whether or not an attack is underway. Interpreting cybersecurity as a sensor fusion problem, includes a number of additional alternative techniques into the solution space. The concept of evidence accrual is to gather measurements over time from different sensors to provide estimates of what event is occurring. A classification fusion technique using feature extraction and fuzzy logic known as Feature Object Extraction is developed and applied to problems such as cyber security and GPS attacks. The feature-aided object extraction technique was developed for the classification problem to fuse different features and generate both a classification and a measure of the quality of the classification estimate. A primary advantage of this is that it evidence is built for each possibility without excluding classes. Thus, the evidence may point to multiple possibilities until evidence disproves a class. Most probabilistic techniques increase the probability of one class by lowering the probability on other classes. Another difference exists in the fact that evidence can be applied to individual classes and not all classes. Feature Object Extraction also allows for a level of evidence to recover from erroneous negative information which might normally cause elimination of a possibility. These design features of Feature Object Extraction are applied to the cybersecurity problem where multiple attacks might be underway simultaneously.

Speaker Bio
Directions
  • This meeting will be offered virtually -- IEEE WebEx connection information will be provided to all preregistered attendees.  You must pre-register on VTools via the link
  • Co-sponsored by San Diego Chapter of the Aerospace Electronic Systems Society
Agenda

5:50 Virtual Networking with AESS Chapter

6:00 Speakers and Program

RSVP

Registration is via vtools: https://events.vtools.ieee.org/event/register/232621

See vtools event notice for more information 



   Home | Section Information | Societies | Announcements | Upcoming Events | Contact

Copyright © 2006–2020 Yellow Interactive, Inc.