Le, Q., and Mikolov, T. 2014. Distributed representations of sentences and documents. In ICML 2014.

https://cs.stanford.edu/~quocle/paragraph_vector.pdf

University of Michigan - Ann Arbor

Before training the prerequisite chain models, the authors extracted dense vector topic representations given sentences or documents. Then used the trained model with a given concept to infer the vector representation that will be used as input to the prerequisite relation learning models. 
- Model: unsupervised Doc2Vec model (Le and Mikolov 2014) using Gensim
- Manner: Distributed Memory Model of Paragraph Vectors (PVDM)
- The dimension of the document vector: 300
- Corpora: LectureBank, TutorialBank, LectureBank+TutorialBank

Learn Before

Related