$$Events$$

Mar. 21, 2017
10:00
-14:00

Room 202 in building 37

On March 21 the Members of PayPal’s Global Risk Data Sciences Club will be holding a joint seminar with BGU researchers working related areas. The goal of the seminar is to strengthen the relationships between PayPal and BGU following similar collaboration with Stanford University, University of California at Berkeley, University of California at Santa Cruz,  Fudan University at Shanghai. PayPal would like to use this opportunity to identify potential research project that PayPal and BGU can collaborate on in the near future. The seminar will consist of three talks by PayPal researchers and a number of talks by BGU researchers. If you are interested in given a talk about your research area, please send a short abstract to DL-PayPal-BGU-Seminar-Abstracts@paypal.com The PayPal team consists of about 10 core members – all with advanced degrees and expertise in data science, machine learning and platform engineering. Our Team provides advanced feature engineering solutions and builds models using advanced machine learning algorithms such as Gradient Boosting, Deep Learning, and Natural Language Processing. We employ big data platforms such as Hadoop and Spark. The primary domain we are focused on is payment fraud prevention. We are also taking responsibilities in other domains such as consumer & merchant marketing and credit​.

 ·        9:00-9:30 gathering and networking
 
 ·        9:30-11:00 PayPal’s talks
 
 ·        11:00-12:00 Ronen Brafman - AI planning -- an overview and
one application
 
 ·        12:00-13:00 lunch break
 
 ·        13:00-14:00 Sivan Sabato - Active Nearest-Neighbor Learning
in Metric Spaces
 
 ·        14:00-15:00 Shimony Eyal - Submodularity and meta-reasoning
in Monte-Carlo Tree Search
 
 ·        15:00-15:30 Coffee break
 
 ·        15:30-15:50 – Chen Avin - Infrence in Social Networks
 
 ·        15:50-16:10 – Hadar Polad – attack graph obfuscation.
 
 ·        16:10-16:30 – Tal Baumel- Sequence To Sequence Attention
Models In PyCNN