About

Ahmed Ashour is a Research Software Development Engineer at Microsoft Advanced Technology Lab in Cairo (ATL Cairo), Egypt.

Prior to graduating college, Ashour joined ATLC as an intern during the two month long ATLC Summer Internship Program 2011, where he worked on adding Location Based Reminders to Windows Phone 7 and Outlook.

After graduating college, Ashour joined ATLC as a CSG on November 2012, but was converted to an FTE on January 2015. During his time at ATLC, Ashour worked on different projects at different fields, but his major fields of focus were Data Mining, Natural Language Processing, Sentiment Analysis, Topic Detection, Keyphrase Extraction.

Ashour is currently working on Sffinx project, which deals with Automatic Relation Extraction from unstructured data, which is to be consumed by Satori.

Other

Projects

During his time in ATLC, Ashour worked on different projects including:

  • Location-Based Reminders on Windows Phone 7 and Outlook.
  • Kernel Canonical Correlation Analysis (KCCA).
  • Scaling Bing Intent classifiers from en-US to ar-EG and zh-CN.
  • Implementing Bootstrapped Classification and Latent SVM.
  • Feature Engineering of the state-of-the-art Sentiment Classification.
  • Implementing the Sentiment Classification Library used by Azure Machine Learning (Azure ML).
  • Using Skip-Gram models and ESA for Topic Detection.

Publications

  • Domain Adaptation for Cross-Lingual Query Classification Using Search Query Logs and Document Classification, by Mohamed Abdel-Hady, Ahmed Ashour, and Rania Ibrahim. Published in The International Joint Conference on Neural Networks (IJCNN) by IEEE. (DOI: 10.1109/IJCNN.2013.6707061)
  • Cross-Lingual Twitter Polarity Detection via Projection Across Word-Aligned Corpora, by Mohamed Abdel-Hady, Riham Mansour, and Ahmed Ashour. Published in ICML WISDOM 2014 Conference. (Link: http://sentic.net/wisdom/2014/mansour.pdf)
  • Feature Selection for Twitter Sentiment Analysis: An Experimental Study, by Riham Mansour, Mohamed Abdel-Hady, Ahmed Ashour, and Hani Amr. Accepted by the Conference on Intelligent text processing and Computational Linguistics 2015 (CICLing)

Patents