Research&Projects

Web Search Engine

University of Pennsylvania, Mar. 2022 - May. 2022 Supervised by Zachary G. Ives

Built a distributed crawler and crawled more than 300000 online documents.
Created customized Hadoop MapReduce jobs to build forward and inverted indices in DynamoDB based on the crawled documents
Created customized Hadoop MapReduce jobs that applied page rank algorithm to calculate the page rank score for each document.
Deployed the search engine server on EC2 nodes which made use of the built indices and page rank score.
Optimized the search logic so that the search query can be finished in fewer than 3 seconds and can take word context into consideration.
Github Link: https://github.com/yx1215/SearchEngine

University of Pennsylvania, Sep. 2021 - Nov. 2021 Supervised by Pratik Chaudhari

Created noisy speech datasets from The LJ Speech Dataset and Urbansound8K Dataset
Modified and applied different neural networks such as WaveCRN, Wave-UNet and GANs for solving the denoise problem.
Achieved to denoise a noisy audio of 0db to 15db with the best model.

University of Pennsylvania, Mar. 2022 - May. 2022

Used React and NodeJS to build a web-based application where user can watch and comment anime.
Made used of the RDS service on Amazon Web Service so that the data can be store on cloud.
Optimized several complex sql queries which originally took more than 20s, and the optimized queries can take fewer than 1s

New York University Shanghai, Feb. 2020 - May. 2020
Supervised by Prof. Olivier Marine

Aimed to create a user-friendly desktop application that allow researchers to conduct large amount of monopoly game simulations within a short time for academic purpose.
Created a monopoly game for simulation and enabled simple AI players using Python, which allows the user to change basic variables such as income and tax in each round. The implemented monopoly game contains most functions in the original monopoly game, such as moving, picking cards, buying/selling properties and trading.
Designed a user-friendly interface using Electron to allow non-CS users to easily run simulations on HPC in NYUSH to improve performance.
Assessed the influence of basic variables on game duration analyzed the reason.
Github Link: https://github.com/Monopoly-Simulation
Download Paper here

New York University Shanghai, Sep. 2019 - Dec. 2019
Supervised by Prof. Lihua Xu

Aimed to build a web-based application for doctor and patients to store medical appointment records.
Used Flask framework in Python to build the database and develop the app backend
Used Bootstrap framework along with JavaScript to develop the app front end.
With the application, patients and doctors will be able to manage their appointments and have necessary conversations via a built-in chat system.
Github Link: https://github.com/yx1215/EHR

New York University, May. 2019 - Aug. 2019
Supervised by Prof. Richard Cole

Aimed to improve a more fair and efficient algorithm for Law School’s course selection system of New York University
Read five papers to understand the advantages and disadvantages of existing resource allocation algorithms
Designed the reasonable algorithms for Law School of New York University
According to the new algorithm designed to establish the applicable data structure to reduce time complexity

New York University Shanghai, Apr. 2018 - May. 2018
Supervised by Prof. Zheng Zhang

Aimed to writing an entertainment platform for multi-person online chat simulation
Completed the chat system based on the simple framework provided to achieve multi-person simultaneous online and group chat
Added the functions of five chess to the chat system, and also wrote a simple AI to support both man-machine and man-to-man battle
Github Link: https://github.com/yx1215/chat_system_with_go_bang