HU Zuying's CV
latest update: Augst 11 2020
- University of California, Santa Barbara (UC Santa Barbara), 09.2019--present
- - Master of Computer Science
- The Hong Kong University of Science and Technology (HKUST), 09.2017--11.2018
- -Master of Science in Big Data Technology
- The Chinese University of Hong Kong(CUHK), 09.2015--07.2017
- -Bachelor of Engineering in Informatioin Engineering
- Top Students Award for MSc BDT Program 2017-2018, HKUST
- School of Engineering Excellent Student Scholarship, HKUST
- School of Engineering Entrance Scholarship, HKUST
- Head's List (Merit) 2015-2016, 2016-2017 of New Asia College, CUHK
- Dean's List 2015-2016, 2016-2017 of Engineering Faculty, CUHK
- Third Class Scholarship for Excellent Students of Sun Yat-sen University
- Course Projects of CV
- Opinion Mining, Independent Project, 09.2017--12.2017
- Mining topics from course reviews
- Parsing the reviews by analyzing dependency relations between words in reviews to get dependency parse tree
- Utilizing double propagation to get opinion words and target words(topics) from reviews with applying extraction rules on dependency tree
- Pruning topics by setting frequency threshold and generating topic phrases by using compound relation between words
- FP-Growth Algorithm implementation based on Spark Framework, Course Project, 10.2017--11.2017
- Implementing of FP-Growth Algorithm in serial.
- Implementing of FP-Growth Algorithm in parallel based on the Framework of Spark
- Analyzing the speed-up of parallel version compared with serial version
- Food Recognition, Course Project, 10.2017--11.2017
- Recognizing Food based on pictures and textual recipes.
- Combining word embedding and tf-idf to represent food recipes in form of vector and to do
dimension reduction, and then applying Multilayer Perceptron to do classification over food recipes.
- Tunning pre-trained Resnet-152 on food images according to the transfer learning,
and then using the tunned CNN to do classification over food images.
- Combining the results from recipe classification and from image
classification based on the logic of ensemble method to improve classification accuracy.
- Image Caption Generator, Undergraduate Thesis Project, 07.2016--05.2017
- Generating a descriptive sentence to the given image.
- Implementing the Image caption generator by combining Neural Networks
including Long-short Term Memory Network(LSTM) and Convolutional Neural Network
(CNN) via torch based on language Lua.
- Researching on the improvement of the Image Caption Generator by
adding attention mechanism in the Recurrent Neural Networks part.