Yun-Shiuan (Sean) Chuang

Master's Student in Computer Science &
Ph.D. Student in Cognitive Science

University of Wisconsin-Madison

About Me

My research focuses on using machine learning to improve human-machine interactions.

I am currently an M.S. Computer Science and Ph.D. Cognitive Science student at University of Wisconsin - Madison. Before that, I was a full-time research associate at Center for Artificial Intelligence and Advanced Robotics conducting research on social cognition module for assistive robots.

My expertise includes data science, AI, ML, conducting research on human-machine interaction.

I am looking for 2021 summer internship as a data scientist, a machine learning engineer, an applied scientist, or a software engineer!

Education

University of Wisconsin - Madison

M.S. in Computer Science & Ph.D. in Cognition and Cognitive Neuroscience (09/2019-present)

  • GPA : 4.00/4.00
  • Selected Courses: Artificial Intelligence, Machine Learning, Deep Learning, Data Structures and Algorithms, Computational Cognitive Sciences

University of Wisconsin - Madison

Visiting Student Researcher (09/2017-06/2018)

  • GPA: 4.00/4.00
  • Fully-funded by Study Abroad Scholarship for Future Scholars (Ministry of Education, Taiwan)

National Taiwan University

Bachelor of Science in Psychology (09/2013-06/2018)

  • GPA : 4.20/4.30 (Rank: 2 / 62)
  • Dean's Award of College of Science; Presidential Awards (7 out of 8 semesters)
  • Selected Courses: Data Science, Statistical Analysis, Psychology Experiment Design, Multivariate Analysis

Skills

Programming Languages

  • Machine Learning: TensorFlow, Scikit-learn
  • Languages: Python, R, MATLAB, SQL, Java, Shell, Git, JavaScript, HTML, CSS, LaTeX

Package Development

  • Author of the R package: label4MRI - a toolbox for labeling brain image (27 stars and 13 forks on GitHub)
  • Contributor of the MATLAB toolbox: rsatoolbox

Advanced Statistics

  • Generalized Linear Mixed Model (GLMM), Markov chain Monte Carlo (MCMC) Sampling, Bayesian Statistics, Bootstrapping & Permutation Testing, Time-Series Analysis, Multivariate Analysis

Research Publications

Machine-Learning Related Publication

1. Chuang, Y.S., Zhou, X., Ma, M., Ho, M. Austerweil, J., Zhu, X. (under review) Using Machine Teaching to Investigate Human Assumptions when Teaching Reinforcement Learners

2. Chuang, Y.S., Hubbard, E., Austerweil, J. (2020) The “Fraction Sense” Emerges from a Deep Convolutional Neural Network. Proceedings of the 42nd Annual Meeting of the Cognitive Science Society [pdf]

3. Chuang, Y.S., Hung, Gamborino E., Goh, O.S., Huang, T.R., Chang, Y.L., Yeh, S.L., Fu,L.C. (2020) Using Machine Theory of Mind to Learn Agent Social Network Structures from Observed Interactive Behaviors with Targets. IEEE, Robot and Human Interactive Communication [pdf]

Cognitive Neuroscience-Related Publication

1. Chuang, Y.S., Austerweil, J., Hubbard, E. (2020). Fraction neural representation in the human parietal cortex: mixed format and magnitude dependent representations. Manuscript in preparation.

2. Chuang, Y.S., Su, Y.S., & Goh, O.S. (under review). Neural responses reveal associations between personal values and value-based decisions.

3. Binzak, J.V., Park, Y., Toomarian, E.Y., Kalra, P., Chuang, Y.S., Percival G.M., P.G., & Hubbard, E.M. (2018) Are Fractions Percepts? Neurocognitive Relationships between Nonsymbolic and Symbolic Ratio Processing in Children and Adults. Presented at Cognitive Neuroscience Society Annual Meeting, Boston, USA [Poster]

4. Chuang, Y.S., Su, Y.S., & Goh, O.S. (2017). Personal core values modulate risky choice evaluation and subsequent risk taking behavior: an fMRI study. Presented at the Society for Neuroscience Annual Meeting, Washington, DC, USA [Poster] [Abstract]

5. Chuang, Y.S. & Goh, O.S. (2017). Personal values modulate risk-taking behaviors: Hedonism as an Accelerator and Security as a Break. Presented at Joint Conference of National Taiwan University, Peking University, and Chinese University of Hong Kong Psychology Departments, Beijing, China [Poster] Received the Excellent Poster Award

Professional Experiences

Graduate Researcher

Computational Cognitive Science Lab, UW-Madison

09/2019-present
  • Applied deep reinforcement learning (DRL) to reverse-engineer human assumptions when teaching reinforcement learners (Python) and found that human is most capable of teaching Q learner with small discount factor
  • Built the online teaching game as a full-stack web developer (HTML, CSS, JavaScript, SQL) for the RL project and collected 1030 players’ data on MTurk
  • Discovered that “number sense” emerged from a deep convolutional neural network (VGG16) trained for classifying natural images (Python, TensorFlow, MATLAB, R) See Publication

Research Associate

Center for Artificial Intelligence and Advanced Robotics, NTU

12/2018-08/2019
  • Developed a robotic social cognition module by training a deep neural network (resnet & LSTM) that infers people’s social preferences with 80+% accuracy (Python, TensorFlow, R)See Publication
  • Designed the online social interaction game for the robotic study as a full-stack web developer (JavaScript, SQL)

Visiting Researcher

Waisman Center, UW-Madison

08/2017-06/2018
  • Applied Multi-Voxel Pattern Analysis - Representational Similarity Analysis (MVPA-RSA) to decode the neural representation of fraction number with human fMRI data See Details
  • Automated the end-to-end human fMRI data (time-series 3D brain images) processing and analysis pipeline for 250+ scans, which slashed 200+ hours of manual work of the lab (Shell, MATLAB)

Research Assistant

Brain and Mind Lab, NTU

10/ 2014-08/ 2017
  • Evaluated the role of personal values in risk-taking behaviors with both behavioral and time-series fMRI data using GLM, functional connectivity, and graph theory measures
  • Discovered that hedonism serves as an accelerator and security as a brake in risk-taking and identified the neural correlates See Details


Data Science Project Leadership

Co-Founder, PyData Madison

12/2019-present

Project Leader at the NTU CS+X Hackathon

Awarded the 1st place for best overall quality among 30+ competing projects

09/ 2016-01/2017
  • Identified the political “political filter bulbbles” that existed in the Facebook user community using k-means clustering (R, Shiny) on web-crawled (JavaScript) large-scale post-liking data (19 million+ observations & 342,796 users)
  • Demonstrated that filter bubbles are formed by users' stances on specific social issues rather than on their support for major political parties See Details

Research Projects

Identification and Examination Of "Facebook Filter Bubbles"

Research Question: On Facebook, do people discuss politics mostly with those within the same filter bubble, i.e., who share the same political stances?

link

Click for More Details

FB filter bubbles

Fraction Neural Representation In The Human Parietal Cortex

Research Question: In the human brain, how are "fraction numbers" encoded?

link

Click for More Details

LAMBDA project overview

Neural responses reveal associations between personal values and decision-making

Research Question: How personal value (i.e., the values you appreciate in life) modulates our decision-making in our brain.

link

Click for More Details

LAMBDA project overview

Other experiences

Contact