When: Monday, September 25, 2017
Where: PGH 563
Time: 11:00 AM – 12:30 PM

Detecting Sockpuppets in Deceptive Opinion Spam

Speaker: Marjan Hosseinia

The research explores the problem of sockpuppet detection in deceptive opinion spam using authorship attribution and verification approaches. Two methods are explored. The first is a feature subsampling scheme that uses the KL-Divergence on stylistic language models of an author to find discriminative features. The second is a transduction scheme, spy induction that leverages the diversity of authors in the unlabeled test set.

Bio:

Marjan Hosseinia is a Ph.D. candidate with particular interests in Text Mining, Natural Language Processing and Machine Learning. Prior to enrolling at the University of Houston, she worked for three years as a software engineer at Regional Information Center for Science and Technology in Iran. She holds a master’s degree in algorithms and computation from the University of Tehran.

Sequence-to-Sequence and Question Answering

Speaker: Fan Yang

Existing QA works, such as machine comprehension QA and community QA, either assume pre-defined reference or rank similar answered questions. We propose to build an intelligent QA system, where references are first retrieved automatically and then Seq2Seq is extended to generate answers as natural language.

Bio:

Fan Yang is a third-year PhD student, advised by Dr. Arjun Mukherjee. His current research interests involve natural language understanding, natural language generation, open domain question answering and abstractive summarization.