Lab Home | Phone | Search
	Center for Nonlinear Studies
	Home  ▶ People   Current   Executive Committee   Postdocs   Visitors   Students   Research   Publications  ▶ Conferences   Workshops   Sponsorship  ▶ Talks   Seminars   Postdoc Seminars Archive   Quantum Lunch   Quantum Lunch Archive   P/T Colloquia   Archive   Ulam Scholar     Postdoc Nominations  ▶ Student Requests   Student Program  ▶ Visitor Requests   Description   Past Visitors  ▶ Services   General     History of CNLS     Maps, Directions   CNLS Office   T-Division   LANL				Thursday, June 23, 20162:30 PM - 2:40 PMCNLS Conference Room (TA-3, Bldg 1690) Student Seminar A New Approach for Anomaly Detection: The Kernelization of Khachiyan's Algorithm Guen GrosklosUtah State University In its standard form, the Khachiyan algorithm efficiently identifies the ellipsoid that encloses every point in a d-dimensional dataset -- it turns out that this is a convex optimization problem. By kernelizing this algorithm, we will be able to develop a more data-adaptive anomaly detector that pays more attention to the periphery of the data while also more effectively enclosing non-Gaussian and non-ellipsoidal data. This will be compared to the kernelized RX algorithm, a well known anomaly detector commonly used for anomaly detection in hyperspectral imaging. Host: Chris Neale