AQFC2015

Keyword Targeting through Deep Learning

----------------------------------------------------------------------------------------------------------------------
 
 
      Department of Systems Engineering and Engineering Management
                             The Chinese University of Hong Kong
 
----------------------------------------------------------------------------------------------------------------------
 
Date: 2:00pm to 3:00pm, May 19, 2017 (Friday)
 
Title: Keyword Targeting through Deep Learning
 
Speaker: Dr. Wen ZHANG, Senior Algorithm Engineer, Baidu Inc.
 
Abstract:
 
Bidword rewriting from query is essential in advertisement targeting system. Traditional approach of keyword targeting relies on the set of alignment rules that are mined statistically from raw data. The corresponding bidwords are retrieved indirectly by rewriting their synonyms in the query. This process, however, have a few shortcomings. First, since the mapping process involves many steps of retrieval and matching, loss of semantics would occur inevitably in this long process of semantic transmission. Second, statistical methods have intrinsic limitations in terms of how much implicit information can be extracted as alignment rules from natural language. Last but not the least, facing enormous amount of data, the expansion of rule sets makes it increasingly difficult to update system in the big data era. In this talk, we discuss the deep learning approach to keyword targeting where targeting is seen as a mapping from one sequence (query can be seen as a term sequence) to another sequence (bidword and advertisement keywords also can be seen as term sequences). Using deep learning approach, a wide spectrum of product lines can accommodated to meet the different requirements of customers.
 
Biography:
 
Dr. Zhang is a senior algorithm engineer in the Fengchao department at Baidu Inc., specializing in pattern recognition and big data mining. She is responsible for developing bidwords classification algorithms used in the firm’s advertisement business. Prior to Baidu, she was an algorithm scientist at 3M cogent where she is responsible for developing production algorithms of fingerprint/voiceprint/face/signature identifications. Dr. Zhang has won numerous awards, including second place in the 3rd fingerprint liveness detection competition (2013), top 3 in NIST fingerprint competition, top3 in FVS competition, and the third place in NIST voiceprint vector competition. Dr. Zhang holds a Ph.D and B.S. in Computer Science from Nankai University.
 
This talk is hosted by Prof. Qi Wu.
 
Everyone is welcome to attend the talk!
 
 
Venue: Room 513,
      William M.W. Mong Engineering Building (ERB),
      The Chinese University of Hong Kong.
 
SEEM-5202 Website: http://seminar.se.cuhk.edu.hk
Date: 
Friday, May 19, 2017 - 06:00 to 07:00