James (Jinhui) Yang

300 North College St, Northfield | (952)495-7367 | yangj2@carleton.edu

Student at Carleton College

I am a quick learner, creative thinker and college student who aspires to venture into the world of programming, especially related to its use of dissecting our day-to-day languages. My interests are in Natural Language Processing and its applications in Digital Humanities, as well as broader applications of machine learning research.

Contact me

Research/Class Projects

Email Subject Generation

Carleton College • October, 2018 — November, 2018

Final project for CS 352 (Computational Models of Cognition) at Carleton College. Built several Long Short-Term Memory architectures using TensorFlow and PyTorch for email subject generation given body text. Work cultimated in a research paper and a class presentation. View it at this link.

Quantifying Emotional Content of Language in Digital Culture

Carleton College • November, 2017 — March, 2018

Currently working on a digital humanities project with Prof. Eric Alexander on the language use of social media story telling. Perform N-gram analysis, topic modeling, sentiment analysis, etc. on the dataset to extract interesting differentiations between posts on social media. View it at this link.

Scheme interpreter in C

Carleton College • October, 2017 — November, 2017

Term project for CS 251 (Programming Languages) at Carleton College. Built a Scheme interpreter in C that tokenizes, parses and evaluates the input Scheme file and prints the result on the terminal. View it at this link.

Tetris game

Carleton College • November, 2017

Final project for CS 257 (Software Design) at Carleton College. Implemented a Tetris game using IntelliJ and MVC (Model-View-Controller) pattern. View it at this link.

Research Experience

Carleton College CS Department

Research Assistant • July, 2018 — August, 2018

I have been working on topic model diagnostics and evaluation using both visual and metric based approaches with Dr. Alexander. We developed a suite of visualizations to effectively explore the parameter space of topic models, and are aiming to extend the results to other variations of LDA topic models such as the Joint Sentiment Model.

University of Minnesota Biostatistics

Research Assistant • June, 2017 — September, 2017

I have been working since Summer 2017 to convert the methods used in a paper by Chu et al. (2003) into a working R package to be published at Journal of Statistical Software.

Developed an R package "BayesSenMC". Performed all necessary commenting and testing. View it at this link.
Wrote and edited a JSS-formatted paper on the R package

Skills

Proficiency in Machine Learning and Natural Language Processing softwares and methods

Extensive experience with NLP libraries/packages (Mallet, Python’s NLTK, and Stanford CoreNLP)
Competent in major Machine Learning models/frameworks (Logistic Regression, KNN, Random Forest, Support Vector Machines, Naïve Bayes, Hierarchal Bayes Model, Recurrent/Convolution Neural Net, Expectation Maximization, etc.)

Proficiency in programming languages

High proficiency in R (ggplot2, dplyr, rstan), Stan, Python (scikit-learn, SciPy, pandas, NumPy, TensorFlow, PyTorch, matplotlib, psycopg2 and Flask), Java, Scheme, C and LaTex
Medium proficiency in HTML, CSS, SQL and JavaScript (D3.js)

Attention to detail

I am a careful programmer as well as a thorough debugger. I always double check the previous line of code while programming. In addition, when there are bugs in a large program, I often trace through each step of the call on paper to make sure all the codes and logic are accurate.

Thorough tester

I am good at finding edge cases as well as creating throrough and clear test files. I have also had experience with unit testing, and can write programs using Java frameworks such as JUnit.