2015
“Machine Learning and Signal processing Methods in Live Business Intelligence Operations” October 15
October 8, 2015
0

IEEE Austin COMSOC/SP Meeting

Machine Learning and Signal processing Methods in Live Business Intelligence Operations

Speakers: Dr. Choudur K. Lakshminarayan

Principal research Scientist in HP

Oct. 15, 2015, 6:00–8:00 pm

Abstract

As the ability to collect data from sensors in industrial applications is increasing at a fast rate, there is a need for algorithms to process data in near real-time to generate actionable insights at the speed of business. To detect patterns in the massive amount of data requires fast, on-line algorithms. Live BI monitors the state of the system, detects a diversity of anomalous events and warns about impending failures. In this talk, I present algorithms using signal processing, non-linear dynamical systems, and time series in a framework known as the mixture of experts to quickly detect changes in patterns in the streaming data. We demonstrate the usefulness of our approach through a problem from the Energy Industry. Our algorithms constructed to detect changes in flow patterns in oil from a production bore-well in real time however are not restricted to oil flows and can easily be applied to other areas such as event tracking in data centers, IT operations, and other applications where large volumes of streaming data is the norm.

Presenters/Bios

Choudur K. Lakshminarayan is a Principal research Scientist in HP. He works in the areas of mathematical statistics, applied mathematics, and machine learning with applications in data mining, data compression, and data analysis of structured and unstructured data in real-time settings. He holds a Ph.D. in Mathematical Sciences and lives in Austin, Texas.

Location:

ATT Labs
Room #220
9505 Arboretum, Austin
Austin,  Texas
United States 78729
Click here for Map

THIS EVENT IS FREE.  Please RSVP at:  https://meetings.vtools.ieee.org/m/36355

For any questions, please contact COMSOC/SP Chair Fawzi Behmann at f.behmann@ieee.org