نبذة عني
Research Scientist with a strong background in Mathematics, Computer Science and Cognitive Science. Expertise includes Information Retrieval, Applied Machine Learning, and Data Mining, with experience developing machine …
Research Scientist with a strong background in Mathematics, Computer Science and Cognitive Science. Expertise includes Information Retrieval, Applied Machine Learning, and Data Mining, with experience developing machine learning algorithms and features for search ranking systems and abuse detection. Skilled in a range of machine learning models and tools including Tensorflow, Pytorch, Hadoop ecosystem technologies, and multiple programming languages.
الخبرة
Research Scientist
Research and development of machine learning algorithms and features for search ranking systems and abuse detection, Training/testing/deploying ranking functions for Y! Japan (Web, mobile & verticals), Training/testing/deploying classifiers for spam account detection & auction fraud detection, Research and development of a search assist system for Y! Japan (spam/gibberish query filtering, adult query filtering & query suggestion ranking), Research and development of hateful/unpleasant question/answer detection system for Yahoo! Japan Answers., Research and development of unrelated question/best-answer pair detection system for Yahoo! Japan Answers, Research and development of adult/spam detection system for Yahoo! Taiwan Answers, Research and development of bidding strategies on the advertiser's side., Research and development of Search Assist Personalization, Research and development of Query Intent Classifier for Shopping/Travel Ads, Research and development of Automated Ad Title Generation from Product Description Text
Technical Expert, Advanced System Division
Research and development of algorithms for satellite/remote-sensing image analysis/enhancement., Development of evaluation tools for the “goodness” of images., Search for a set of image measurements and an appropriate function of those measures that correlates the ranking of images by human observers.
Postdoctoral Researcher & Assistant Specialist (Software R&D)
Research and development of the breast cancer tissue classification system as a new feature for the image analysis system that was previously developed and implemented., The system, implemented in C++, utilizes neighborhood level information of each cell nucleus and minimal spanning tree statistics to gather/classify structural information of each tissue., Design and implementation of unsupervised texture segmentation algorithms in C++ to enhance the classification capability of the breast cancer tissue classification system., Design and implementation of the on-line EM algorithm in C++ and Java.
المشاريع
Ad-group assignment to keyword
Lead the project following "ad-group assignment as a search" framework. Use both the relevance matching (search index building) and the semantic matching (sentence transformer) to achieve the satisfactory level of precision and recall. The ad sales team has decided to use the proposed method.