Student research opportunities
Extracting User profile with Deep Learning Techniques
Project Code: CECS_1123
This project is available at the following levels:
CS single semester, Honours, Masters, PhD
Supervisor:
Dr Lizhen QuOutline:
Understanding the user behaviour on the web is essential for market analyses and financial trend prediction.
However extracting user profile in a supervised way is expensive due to manual construction of huge amount of training data. Deep learning techniques allow us to use a small amount of labelled data by leveraging large amounts of unlabelled data. Therefore, in this project, we aim to create a cutting edge Deep Learning systems to extract user profile from the web with minimal manual effort.
Requirements/Prerequisites
Familiarity with linear algebra, probability, and natural language processing. Knowledge of online learning and deep learning would be a plus. Good coding skills in Scala or Java.
Background Literature
Jiwei Li, Alan Ritter, Eduard Hovy. Weakly Supervised User Proļ¬le Extraction from Twitter. Baltimore. ACL, 2014
Abel, Fabian, et al. Semantic enrichment of twitter posts for user profile construction on the social web. The Semanic Web: Research and Applications. Springer Berlin Heidelberg, 2011. 375-389.
