A log of work that I have spent some time on
Storage
[HyperPebblesDB] (2018 - ): a key-value store that is part of LevelDB family
[Shuati] (2017 - ): a collection of solutions to OJ problems with the goal to stay fresh with computer science fundamentals and learn new programming languages. Also include a python scraper for solution template formation.
[Emacs Config] (2016 - ): my personal emacs configuration that I used during the day work. Very lightweight.
[Algo] (2016 - 2018): a collection of data structures, algorithms with the aim to keep my computer science knowledge fresh. Implemented in C.
[Active Learning for Neural Dependency Parsing] (2018): explore the idea of active Learning on Stanford Neural Dependency Parsing
[Part-of-Speech Tagging with LSTMs] (2018): build a BiLSTMs model with orthographic features
[Distributed Key-Value Store] (2018): build a distributed Key-Value Strore using eventually consistency model with two session guarantees: Read Your Writes and Monotonic Reads
[Bidirectional Bigram Model] (2018): build a bidirectional bigram model
[Identifier Inference through Neural Network] (2017): build N-gram and Neural Network language models to study the identifier naming convention problem
[Investigation of neural network architecture on MNIST digits dataset] (2017): implements a FeedForward Neural Network and a Convolutional Neural Network for the handwritten digits classification task using Keras
[Exploring Stereotypes and Biased Data with the Crowd] (2017): How can we use the crowdsourcing to alleviate the bias in the dataset for machine learning?
[Modular reinforcement learning for 2D navigation] (2017): Use the modular reinforcement learning to explore a 2D grid navigation task
[Gaussian process for human movement data] (2017): Use the gaussian process regression to study the human movement data
[Neural Networks for Sentiment Analysis] (2017): implements a FeedForward Neural Network and a Convolutional Neural Network for Sentiment Analysis using Tensorflow
[Shift-Reduce Parsing] (2017): implements a shift-reduce parser to generate dependency trees using both a greedy model and a global model with beam search (logistic regression, Structured SVM, arc standard transition system, beam search)
[Sequential CRF for NER] (2017): implements HMM model for POS tagging and CRF model for NER (forward-backward algorithm, viterbi algorithm, beam search)
[ICA on mixed signals] (2017): implements ICA to perform blind source separation
[PCA on handwritten digits] (2017): implements PCA on handwritten digits data and run KNN to capture the major features of the images
[OptiTimal] (2012): an android application to allow user to track their time allocation and obtain a statistical report about their time management