Gomoku Project Report
Shining points:
-
Used Temporal-Difference Learning in taking step
-
Threat-space was used for must-win strategy
-
Implemented the Numpy numerical operation on our own with List to compact the size of executive file (.exe)
-
Overpower pisq-7 (ranked 43/58 in published Elo Ratings) with a score of 13:7 and ranked 5/30 in the class
Technique: spark RDD,MLlib Report1 Report2
Wiki with LSA Report
Technique: Spark TF-IDF, SVD
Hadoop Labs Report
HBase Labs Report
Spelling Correction Report
Technique: Kneser-Ney smoothing See my blog article
Sentiment analysis Report
Technique: Word2Vec
Chinese event extraction(Sequence Tagging) Report
Technique: HMM, CRF model
English Corner Design document User document
Technique: UDP based client Chatting system
Shining points:
-
Detailed User Document
-
Diversity mode: Secret and Public, different language's own rooms
-
Drop-down bar auto-completion