Simplicity vs. Complexity in the Construction of Efficient Algorithms
Share this Session:
  Omar Abdala   Omar Abdala
Chief Data Scientist
Lotame
http://www.lotame.com/
 


 

Thursday, April 6, 2017
08:30 AM - 09:15 AM

Level:  Intermediate


The more data you introduce, the shakier those elegant designs become. Often you’re faced with one of two situations: 1. The runtime of elaborate algorithms ultimately makes them incomputable, or 2. The value they bear over basic algorithms progressively diminishes. How do we choose between a highly detailed model or a less detailed model that uses all the data? How do we choose between parametric models and less theoretically appealing non-parametric methods? What’s the impact of the size of web scale data we’re all faced with on these choices? What we find is that sometimes a simple model is better, even if it’s unpalatable at an intellectual level. I will discuss when the latter makes the most sense, and in doing so will touch upon:
  • Monetizing data exhaust and analytics
  • Investing in the big data industry
  • Big data frameworks
  • A rough overview of a few modeling algorithms


Omar Abdala is Chief Data Scientist at Lotame and recent winner of the AMA 4 Under 40 Emerging Leaders Award. He is the author of over 15 technical papers and patents on statistical modeling, inference, ad network optimization, targeting and inventory management. Omar joined Lotame through the acquisition of AdMobius, the first mobile Audience Management Platform, which he co-founded as Chief Scientist. Prior to that, Omar was a Data Scientist at Apple, leading the creation of iAd’s targeting database using data assets across the Apple Corporation. Omar arrived at Apple by way of Quattro Wireless, one of the first mobile ad networks, where he designed their core technologies in ad inventory management, network optimization and analytics. Omar holds a B.S. and M.Eng. in Computer Science and Electrical Engineering from MIT, and is a Ph.D. candidate on leave at the School of Engineering and Applied Sciences at Harvard University.


   
Close Window