140901 - Zhongda Xie

From cslt Wiki
Jump to: navigation, search


1. Review the papers in "semantic matching for IR", classify them into various categories and make comparisons.

2. Make preparations for building a "semantic matching toolkit", which integrates all the methods mentioned in task 1. Make a survey about the necessary data, source code and inference papers.


First implement a popular method: Markov Random Field Model.