Create conversational dialog manager for Discover Deis - a voice helper tool for navigating around Brandeis

A joint work with other classmates that we implemented a dialog manager for Discover Deis with the help of Wit.ai. We wanted to make the software more interactive and helpful. Moreover, it is a step toward an analysis of the aspects of complexity of an application that will have implications for its dialog manager.

Final project of Computational Models of Discourse class of 2016

Vietnamese syntactic parser with Haskell

Implemented a parser in Haskell that can syntactically parse a shortened version of Vietnamese Cinderella fairy tale.

Final project of Computational Semantics class of 2014


Vietnamese full text recovery from diacritic-lacking documents with abbreviations

Used Hidden Markov Model (HMM) to recover Vietnamese text from a simplified form (abbreviations and lack of diacritics) to full form. This is an important task in processing social network comments and updates, which use an impoverished form of Vietnamese.

Final project of Machine Translation class of 2014


Location entity detection and categorization for Bacteria Biotope (BB) task with Nested Entities Detection (LBBN)

Used Conditional Random Field to find extents of all habitats of any bacteria species or bacteria strain in the document using BIO tags (Beginning, Inside, Outside). I used a similar method to find subentities of generated entity candidates, and add them into the pool of entities.

2013 BioNLP Shared Tasks, Final project of Information Extraction of 2013


PartML: meronymic relationship annotation and classification by machine learning

This project created an annotated resource for the task of extracting part-whole relationship in text. Example of this relationship is: in I attend Brandeis, I myself is a part/member of Brandeis as a whole.

Formularizing the problem as named entity relationship classification, using Weka, we compared a set of ML methods, and found that Naive Bayes actually works the best for the task.

Term project of Natural Language Annotation for Machine Learning class of 2013