School of Computer Science · Master of Information and Technology Strategy

School of Computer Science · Bachelor of Science

Python · Java · C · JavaScript · C++ · Bash · Assembly · Scala · HTML · CSS · R · PHP

Git · AWS · Node.js · Express · Django · TensorFlow · Scikit-Learn · Java RMI · RESTful · Elastic Stack · Splunk · MySQL · NoSQL · MATLAB · Linux · LaTeX

Machine Learning · Deep Learning · Distributed System · Cloud Computing · Algorithms

English (Professional Working Proficiency) · Chinese (Native) · Japanese (Elementary proficiency)

FL Studio · Ableton Live · Analog Lab Lite

Pittsburgh, PA · Software Development Intern · Advisor: Hasan Yasar · View Project

Shanghai · Research and Development Intern · View Project

Shanghai · DevOps Research and Development Intern · View Project

Waterloo, ON · Research Intern · Advisor: Ming Li · View Project

Irvine, CA · Research Intern · Advisor: Chen Li · View Project

Boston, MA · Research Assistant · Advisor: Li Zhou · View Publications

YITU Technology

Shanghai · Research and Development Intern · Apr 2018 - Jul 2018

Applied Mixup as a data augmentation method for image classification tasks, increased accuracy by 0.5-0.7% on 28 datasets, reached 96.2% accuracy on Cifar-10 dataset. Added Inception series algorithms into TF Autobot. Accomplished regression tests on TF Autobot. Built a speed testing tool for Caffe, TensorRT model on CPU/GPU respectively in C++.

eBay Engineering and Research Center Co., Ltd

Shanghai · DevOps Research and Development Intern · Apr 2018 - Jul 2018

Revamped an online intelligent operation platform for DevOPS department, used Django as python web-frame work and Vue as JavaScript framework. Built a Scala project of back-end algorithms including unsupervised learning models (Kmeans, Hierarchical Clustering) and abnormal detection on the result of each pool (server) parameter's feature extraction, validated the risk level for each pool based on the classified history data, applying Apache Spark for distributed computing.

Splunk Application for Medical Device Vulnerability Detection and Data Enrichment

Carnegie Mellon University · Capstone Project · Sponsored by Highmark Inc. · Mar 2019 – Present

Developed a customized Splunk application for medical device vulnerability search and aggregation of Rapid7 Nexpose records, hospital records, Mitre and NVD database using Javascript, Node.js, and XML. Built Splunk backend and node server using Splunk API, applied unit test using Artillery, performance test using Splunk REST API, and workflow test using Selenium.

READ MORE

AWS Based Data Intensive Elastic Distributed Cloud Server

Carnegie Mellon University · June 2019 – Present

Designed and built a 4-tier, auto-scaling, e-commerce web service using Node.js with Express, indexed MySQL database using AWS RDS. Deployed on AWS EC2 with ELB (Elastic Load Balancer). Implemented central queueing, optimized on multiple load-balancing method.

READ MORE

QryEval: Text-Based Search Engine

Carnegie Mellon University · Jan 2019 – Apr 2019

Project for CMU course 11-642 Search Engine. Built a text-search engine with Apache Lucene in Java, containing a corpus of 500,000+ documents. Implemented multiple retrieval models including Ranked/Unranked Boolean, Okapi BM25, Indri, etc. Added SVM re-ranking, query expansion, intent-aware diversified search, and forward indexing as additional features.

READ MORE

Group Photo Collage: Distributed Transaction System

Carnegie Mellon University · Mar 2019 – Apr 2019

Project for CMU course 15-640 Distributed Systems. Implemented a distributed transaction system using two-phase commit using Java. Dealt with lost and delayed messages. Handled node crashes and node recovery. Utilized thread-safe data structures and locks to process concurrent requests, used logging to persistent storage for failure recovery.

READ MORE

File Caching Distributed System and Remote File Operations System

Carnegie Mellon University · Jan 2019 – Mar 2019

Projects for CMU course 15-640 Distributed Systems. Built an RPC system to allow concurrent remote file operations in C. Designed and implemented a distributed system for file caching using Java RMI and Java threading. Designed and built a caching proxy, robustly handled multiple concurrent clients and ensured open-close session semantics on concurrent file access.

READ MORE

Rotational Dropout – A domain Adaptation Method for CNN Architectures

Carneigie Mellon University · Sep 2018 - Dec 2018

Project for CMU course 10-805 Machine Learning for Large Datasets (PhD Level). Integrated the merit from both Dropout and Batch Normalization for each building block by keeping consistent variance, brought improvements on state-of-art CNN networks by efficiently mitigating overfitting problems. Achieved 4.74% error rate on CIFAR-10 using ResNet50 and 4.56% error rate using DenseNet.

READ MORE

Concurrent Caching Web Proxy

Carneigie Mellon University · Nov 2018 - Dec 2018

Project for CMU course 15-513 Introduction to Computer Systems. Designed and created a sequential caching web proxy in C, which serves for both local server and public websites. Upgraded the proxy to deal with multiple concurrent connections using POSIX Threads, caching both static and dynamic contents using LRU eviction policy and dealt with race conditions and synchronization using blocking.

READ MORE

Malloc Lab: Dynamic Storage Allocator

Carneigie Mellon University · Oct 2018 - Dec 2018

Project for CMU course 15-513 Introduction to Computer Systems. Created a general purpose dynamic storage allocator for C programs which supports calls to malloc, free, realloc, and calloc functions. Use doubly-linked segregated list to preserve free block list, used LIFO as insertion policy.

READ MORE

Linux Shell for Job Control and I/O Redirection

Carneigie Mellon University · Sep 2018 - Oct 2018

Project for CMU course 15-513 Introduction to Computer Systems. Implemented a Linux shell program, tsh (tiny shell), that supports a simple form of job control and I/O redirection using process control and signalling.

READ MORE

Glutton: Full-Featured Web Application

Fudan University · Mar 2017 - May 2017

Project for Introduction to Database Systems (COMP130010.03) @ Fudan University, an full-featured web application for lazy gluttons. Built with Python Flask framework, SQLite, jQuery, HTML5, CSS.

READ MORE

Cloudberry

University of California, Irvine · Advisor: Chen Li · Jul 2017 – Sep 2017

Built part of middleware on top of a big data management system, Apache AsterixDB, mapped front-end query to MySQL and PostgreSQL query in Scala. Wrote the SQL translator and connector of middleware to databases which sent and received queries and transferred datatypes for semi-structured data models. Developed a generalized data structure for back-end translators to support efficient real-time analytics.

READ MORE

De Novo Peptide Sequencing by Deep Learning

University of Waterloo · Advisor: Ming Li · Sep 2017 - Oct 2017

Contributed to the development of DeepNovo, a deep neural network model about de novo peptide sequencing using ion-CNN, spectrum-CNN, and long short term memory (LSTM) recurrent neural network (RNN) in Tensorflow. Promoted the recall accuracy by 6.2% for training model about inferring sequences from a tandem mass spectrum on the analysis of data independent acquisition via parameter adjustment. Used Cython to speed up 50% of data feeding process in Python and decrease training perplexity for 59.4%.

READ MORE

RegionAl: an Optimized Regional Classifier to Predict Mortality in Chronic Obstructive Pulmonary Disease Patients.

Haohan Zhang, Chunlei Tang, Joseph M Plasek, Yun Xiong, Jing Ma, Li Zhou, David W Bates
2018 American Medical Informatics Association Annual Symposium (AMIA)

We previously developed a regional classifier based on a spiral timeline for visualizing literature data, which presents research topic words under different themes in a spiral map to show the chronological development of focused research topics. When timelines are combined with a geographical map, they depict temporal patterns of events with respect to their spatial attributes. we call a “spiral atlas of temporal variation” (abbreviated as “atlas”).

READ MORE

A Deep Learning Approach to Handling Temporal Variation in Chronic Obstructive Pulmonary Disease Progression

Chunlei Tang, Joseph M Plasek, Haohan Zhang, Yun Xiong, David W Bates, Li Zhou
2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

Chronic Obstructive Pulmonary Disease (COPD) is a leading cause of mortality in the United States. Representing COPD progression using temporal graphs may offer critical clinical insights. LSTM (Long-Short Term Memory) units in recurrent neural networks can process data with constant elapsed times between consecutive elements of a sequence but cannot handle irregular time intervals (i.e., segments with unequal-time). In this study, we propose a four-layer deep learning model that utilizes a specially configured recurrent neural network to capture irregular time lapse segments. Experiments on a corpus of COPD patients' clinical notes compared to baseline algorithms showed that our model improved interpretability as well as the accuracy of estimating COPD progression.

READ MORE

Developing a regional classifier to track patient needs in medical literature using spiral timelines on a geographical map

Chunlei Tang, Haohan Zhang, Kenneth H Lai, Yuxuan She, Yun Xiong, Li Zhou
2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

In this paper, we propose a linear classifier based on a spiral, which we call a regional classifier. The study emphasizes the development of visualization methods and the process of finding a specific research clue to track patient needs reported in medical literature. When timelines are combined with a spiral geographical map, they show a geometric shape that helps to reveal the clues from different spatial viewpoints and periodical constraints. Our evaluation showed that the regional classifier produces better visual effects than support vector machine classifiers.

READ MORE