Site Search
Computer Science


Dr. Scott GoodwinDr. Scott Goodwin
Dr. Scott Goodwin
Lambton TowerLambton Tower
Lambton Tower
Arunita Jaekel, Ph.D.Dr. Arunita Jaekel
Dr. Arunita Jaekel
Xiaobu Yuan, Ph.D.Dr. Xiaobu Yuan
Dr. Xiaobu Yuan
Windsor WaterfrontWindsor Waterfront Park
Windsor Waterfront Park
Robin Gras, Ph.D.Dr. Robin Gras
Dr. Robin Gras
Imran Ahmad, Ph.D.Dr. Imran Ahmad
Dr. Imran Ahmad
Christie Ezeife, Ph.D.Dr. Christie Ezeife
Dr. Christie Ezeife
Dr. Luis RuedaDr. Luis Rueda
Dr. Luis Rueda
Alioune Ngom, Ph.D.Dr. Alioune Ngom
Dr. Alioune Ngom
Dr. Robert KentDr. Robert Kent
Dr. Robert Kent
Jessica Chen, Ph.D.Dr. Jessica Chen
Dr. Jessica Chen
Dr. Ziad Kobti lecturingDr. Ziad Kobti
Dr. Ziad Kobti

Learning Embeddings for Academic Papers

Add this event into your calendar using the iCAL format
  • Wed, 06/13/2018 - 1:00pm - 3:00pm

Learning Embeddings for Academic Papers

PhD. Thesis Proposal by:

Yi Zhang

Date: Wednesday, June 13, 2018
Time: 1:00 pm - 3:00 pm
Location: 3105, Lambton Tower

Abstract: Learning vector representations is essential for many downstream tasks such as clustering, classification, regression, information retrieval, and recommendation. Academic papers contain not only plain text, but also hyperlinks such as citations, making them hard to represent.

SGNS (Skip-gram with Negative Sampling) is the state-of-art word embedding algorithm proposed in 2013. In theory, it equals to factorize over word co-occurrence matrix. In practice, SGNS and its variants have been successfully applied to learn high-quality embeddings from documents, networks, and linked-documents.

In our work, we discuss the norm convergence issue of SGNS and propose to use L2 regularization to improve the embeddings. We summarize multiple approaches along with our method to learn the word, network, document, and paper embeddings from academic papers. For word embedding, our method is evaluated in word similarity and analogy benchmarks; for network/document/paper embeddings, our methods are evaluated in classification tasks.

Thesis Committee:     
Internal Reader: Dr. Mehdi Kargar, Dr. Dan Wu
External Reader: Dr. Qingming (Jonathan) Wu (ECE)
Advisor: Dr. Jianguo Lu

See More: