CSE Seminar: Oleg V. Diyankov

*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************

Event Details
  • Date/Time:
    • Friday September 21, 2012 - Saturday September 22, 2012
      2:00 pm - 2:59 pm
  • Location: Klaus 2447
  • Phone:
  • URL:
  • Email:
  • Fee(s):
    N/A
  • Extras:
Contact

Mark Borodovsky, borodovsky@gatech.edu

Summaries

Summary Sentence: Data Analysis and Inference Software Systems for Real Life Problems

Full Summary: No summary paragraph submitted.

Speaker:

Dr. Oleg V. Diyankov
CEO, Neurok Software LLC

Title:

Data Analysis and Inference Software Systems for Real Life Problems 

Abstract:

We have developed several data analysis and inference software systems utilizing classical and Bayesian statistics, machine learning, time series analysis, and experimental design. 

The software systems include among others:

1. Oil pipeline defects recognition and quantification:

Software discovers pipeline wall defects from massive streams (terabytes) of magnetic and ultrasonic sensor readings. 

2. Hand-held, high-explosive detector (nuclear plants security systems):

Software performs real-time, multivariate (100 variables) microwave length data analysis.

Data anomalies represent presence of explosive. The predictive system has high recall and precision.

3. Adaptive performance control of iterative solver of sparse linear algebraic systems:

SparSol – a software library – was developed for solving large sparse linear systems. SparSol implements an ideology of preconditioned iterative methods which are quite effective for applications based on discretizations of PDEs on 3D grids. Software consistently improves run time performance of real FE/FD simulation models, by 30% or more, compared to the performance of models with default linear solver settings.

4. Noisy text analytics: Robust software tools for automatic extraction of information from noisy biased longitudinal unstructured text data (news feeds, email archives, intranet file servers, SQL/ODBC databases, live chat, etc.). 

 

These systems were developed with active participation of Nelli Fedorova, Elena Kartasheva, and Vlad Pravilnikov.

 

Neurok Software LLC is small business created in 2001 by a group of former nuclear scientists who worked together at the Russian Federal Nuclear Center – VNIITF (Snezhinsk, Russia).  Starting with six employees, the company now has 20+ employees in the US office (Alpharetta, Georgia) and 20+ employees in the office in Russia.

Neurok Software’s main area of expertise is development of mathematical algorithms and software.  The company does not typically produce complete applications but rather works to make existing applications faster, more efficient and robust.  Neurok Software is particularly interested in development of algorithms and software for 3D irregular grid construction.

 ------------------------------------------------

  Mark Borodovsky at:    borodovsky@gatech.edu

Additional Information

In Campus Calendar
No
Groups

High Performance Computing (HPC), College of Computing, School of Computer Science, School of Interactive Computing, School of Computational Science and Engineering

Invited Audience
No audiences were selected.
Categories
Seminar/Lecture/Colloquium
Keywords
No keywords were submitted.
Status
  • Created By: Lometa Mitchell
  • Workflow Status: Published
  • Created On: Sep 18, 2012 - 11:02am
  • Last Updated: Oct 7, 2016 - 10:00pm