Minseok "Mason" Jung
Massachusetts Institute of Technology
Massachusetts Institute of Technology
MIT CSAIL
MIT IDSS Technology and Policy
SERC Scholar (22'-23')
I am a graduate student researcher at the Decentralized Information Group (DIG) at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT).
AI Policy and Bias/Fairness in ML are my core research topics with backend-oriented approaches. They are vital concerns for reliability, effectiveness, and humanity in the advancement of computing. To respond to these queries, I am navigating the leading edges while investigating the fundamentals of AI aimed to counteract misuses and lead accountable utilizations.
My current work is initiating AI writing policies for MIT. I believe humans need to use their own words and creativity to symbolize their ideas. I utilize quantitative techniques to inspect, analyze, and predict. I use qualitative approaches to decide, question, and direct.
Dr. Lalana Kagal supervises my endeavors.
I have been honored to be a Social and Ethical Responsibilities of Computing (SERC) Scholar at MIT Schwarzman College of Computing. In 22'-23', 56 SERC Scholars from five schools were selected for cutting-edge computational projects.
Winning the 1st place prize MIT Policy Hackathon among 280 pre-selected participants in 2021 directed me to think about societal aspect of Computer Science. At MIT, I lectured <AI, Systems, and Policy> at the IDSS's Commencement for MITx MicroMasters in Statistics and Data Science.
During my undergraudate, I did research in the Social Computing Lab and Governance Lab on Sociotechnical Systems and delved into back-end systems and data modeling with a focus on compliance and societal responsibilities.
You can check detailed information about me on my CV and Linkedin. I listed major projects that I worked on below.
Projects (n = 10)
[1] Winning the MIT Policy Hackathon 2021
MIT_Hackathon_Certificate.pdfWon 1st place in the MIT Policy Hackathon 2021. My team, Policy4ALL, used social data analysis (e.g., time series, sub-group, etc.) to tackle the Lumen Project's open internet challenge. I featured improving consistency on global activities in policy and terms, and multi-body observation on the internet takedown. I framed and wrote the policy paper for all. The paper was submitted to the Berkman Klein Center.
[2] Mapping Geographical Biases of AI Principles
Analyzed AI principles with the objective of identifying bias on choropleth maps. The result revealed that 89% of policy codes were dominated by three regions. [Paper] or [Conference].
Contributed to the development of [Dynamics of AI Principles] at the AI Ethics Lab by using data and framework.
Supervised by Dr. Madelyn Sanfilippo
[3] AI Principles and Philosophical Foundations of AI
[4] Public AI Ethics and Non-Public AI Ethics
Investigated philosophical foundations of AI principles, especially the value of 'eudaimonia' in IEEE's AI Principles. Compared characteristics of public and private AI ethics. Made compelling questions.
Developed through Computer Science + Philosophy courseworks and research guided by Dr. Kishida Kohei and Prof. Helga Varden.
[5] AI, Decision-Making, and Society
Figuring out that AI detection is biased against groups with over-representation and under-representation using four fairness metrics.
[6] Income and Life Satisfaction: How Are They Related to Personal Background?
Employed advanced statistical analysis to process social data. The work figured out what psychological aspects are relevant to life satisfaction using Midlife in the United States (MIDUS) Series dataset.
From a course: IDS.131
[Problem Set 3] -- data cannot be disclosed
[7] Scalable Community Participation in Urban Design for Social and Ecological Resilience
Developed computational methods that integrated computer-aided design (CAD), Geographic information system (GIS), and co-design approaches through data analysis and field study.
This SERC project contributed to the future plan of the New York City Housing Authority (NYCHA).
[8] AI Writing Policy (In-Progress at DIG)
Developing "AI Writing and Policy: Needs and Effectiveness" for MIT. This is my master's thesis. I employ predictive models to measure its effectiveness and use qualitative approaches to address the needs for AI writing policy.
Supervised by Dr. Lalana Kagal
[9] Biases of Data and Algorithms (In-Progress)
Examining various forms and sources of data and algorithmic bias to understand the multiple factors contributing to AI bias through the conceptual analysis and case studies (about bias 'of', not 'in', data and algorithms). [Paper] and [Presentation]
Would like to integrate more case studies and validate the framework by comparing other papers. I am developing computational simulations for each section.
Supervised by Dr. Sangyoon Yi
[10] Fairness in ML (In-Progress at DIG)
Working on... Almost done!!!
(Image is irrelevant: a manual run of Turing Machine)
[+a] Marathon!!!
Finished Cambridge Half Marathon on November 5th, 2023.
Keep doing training for another half marathon to conquer the full marathon!!!