2015-01-21 Top 20 Data Quality Solutions & Random Walks for Scale Space Theory
University of Colorado Denver - Wednesday January 21, 2015
Top 20 Data Quality Solutions for Data Science - Abstract
Data quality continues to be one of the chief challenges, costs and reasons for project failure in data science. Problems in this space limit accuracy, destroy credibility and can result in harmful solutions. And unlike challenges such as scalability and cost it has seen no major breakthrough improvements. This presentation will cover the types of problems, as well as their impacts, causes and various solutions.
Ken Farmer - Bio
Ken Farmer is the senior data architect/wrangler/librarian for ProtectWise where he is developing their analytical data solution. Previously, he has developed, maintained, managed and consulted on analytical data architectures for IBM, MapQuest, Verizon, and others.
Lost at See: Random Walks for Scale Space Theory in Computer Vision - Abstract
A brief overview of scale space theory and its connection to random walks is given. A tool to find patterns of a particular scale in large data sets is presented. It is shown how modern web applications can assist the process of scale detection.
Florian Sobieczky - Bio
Florian earned his PhD at Goettingen in Germany and post docs at the University of Techology at Graz (Austria), and Berlin (Germany). He was an ULAM fellow in 2011/12 at CU Boulder and is now a lecturer at the University of Denver. His interests include: Probability Theory - Stochastic Processes on Discrete Structures: Random Walks, Percolation, Queueing Theory. Florian has authored the following journal articles: [1] `Bounds for the annealed return probability on large finite percolation clusters', Electronic Journal of Probability, 17 (2012), no. 79, 1-17; and [2] `An interlacing technique for spectra of random walks and its application to finite percolation clusters', Journal of Theoretical Probability, Vol. 23, No. 3, (2010), 639-670.
Marilyn Waldman - Bio
Marilyn is a software engineer focused on full stack API’s. She holds an MS in Computer Science from the University of Colorado and has over 18 years of experience. Aside from code, Marilyn’s interests are in the fields of computational complexity and massively parallel, distributed computing.