Hi there! This is Haoshen Hong, CS Master in Stanford. I recently graduated from UC San Diego with B.S. in Applied Math. I'm interesting in deep learning and its application in NLP and audio data.
Build models to perform feature-level study in CNN using tensorflow.
ArXiv: https://arxiv.org/abs/1712.02400
Semantic analyze and extract information in Chinese court cases in a rule-based approach
Learn some NLP processing techniques, including but not limited to CRF and Viterbi algorithm
Explore SQuAD v1.0 (Reading comprehension Q&A dataset) with a special focus on BiDAF + Self attention model.
Gain some basic understandings of different embeddings (Word2Vec, GloVe, ELMo) and their applications in NLP works.
Learn some pratical approaches in dealing with NLP work (Feature Extraction, RNN architecture, RNN Regularization, etc).
Stacking Xgboost, Lasso and Feedforward Neural Network to predict house price in Python
Score top 5% in public leaderboard (Current rank 13/2287, score 0.11189)