Employed neural networks to do function approximation and parameter estimation of partial differential functions
Wrote python script to find lower and upper bounds and neural importance of feedforward neural networks which fit
Genz integrand families functions and visualized the results
Worked in the auto-grading group and aimed to build model to automatically label the role of sentences, such as topic sentence, in TOEFL writings
Extracted and validated candidate features of sentences in TOEFL writings with experts
Retrained pre-trained Word2Vec and Glove models for sentence embedding by adding our corpus
Built a model by combining bidirectional LSTM and CRF with TensorFlow to conduct sequence tagging on TOEFL
writings and improve the test accuracy to 0.71
Involved into Store Locator project, wrote a web crawler with Python to automatically collect basic information of US retailers and archived the information with a proprietary database
Participated into ShoppingSite project, checked and updated goods information with SQL queries