Harvey Mudd College
Harvey Mudd College

Harvey Mudd College

Harvey Mudd College (HMC) is highly selective undergraduate liberal arts college (850 students) emphasizing science, mathematics, and engineering. HMC located in Claremont, California, is part of the Claremont Colleges, a consortium that includes five colleges and two graduate schools.

Research — Music Information Retrieval

Sheet music-MIDI alignment, Deep Learning, Computer Vision
MIDI-Sheet music Alignment
At the Music Information Retrieval Lab, I developed a dynamic programming algorithm for performing multi-modal alignment between sheet music and its corresponding computer-synthesized MIDI. To accomplish this, my team introduces a new representation (called ‘bootleg’) that can be generated from MIDI. I created a deep fully convolutional network for detecting musical notes on sheet music, used for creating the bootleg representation. We published this work in ISMIR 2019 and presented at Delft, the Netherlands.
Midi-sheet music alignment using bootleg score synthesis Proc. of the International Society for Music Information Retrieval Conference (ISMIR), 2019. PDF: https://arxiv.org/pdf/2004.10345.pdf
MIDI Passage Retrieval
Leveraging the idea of bootleg representation, my team extends the problem to MIDI passage retrieval given the sheet music images taken from a regular cellphone. The biggest challenges are in the differences between digitized sheet music (i.e., in PDF format) and photos of sheet music. I further improved the bootleg representation so that it’s more rigorous to page curvature, lightning, and other deviations. The results are published in ISMIR 2019 and IEEE TMM 2020.
MIDI passage retrieval using cell phone pictures of sheet music Proc. of the International Society for Music Information Retrieval Conference (ISMIR), 2019. PDF: https://arxiv.org/pdf/2004.10347.pdf
Using cell phone pictures of sheet music to retrieve MIDI passages IEEE Transactions on Multimedia 22, no. 12, pp. 3115-3127. 2020. PDF: https://arxiv.org/pdf/2004.11724.pdf

Research — FPGA/Microcontroller-Based Embedded Systems Lab

FPGA, Microcontroller, Embedded Systems
PCB Board for FPGA/Microcontroller Classes
My team designed and developed a PCB consisting of a microcontroller SAM4S and a Cyclone IV FPGA. The PCB is used in a microprocessor-based systems class. Together with the board, I created lab instructions and homework based on this newly developed PCB. The PCB board is publicly available, and the lab instructions are published in GLSVLSI 2020.
A Board and Projects for an FPGA/Microcontroller-Based Embedded Systems Lab Proceedings of the 2020 on Great Lakes Symposium on VLSI, pp. 561-565. 2020. PDF: https://dl.acm.org/doi/abs/10.1145/3386263.3406930

Research — Theoretical Machine Learning Lab

Theoretical Machine Learning
Information Gain in Infinite Space
In many machine learning algorithms, a natural way of measuring the reduction of uncertainty is either taking the proportion of the target space to the total space. The problem arises when the space is unbounded and uncountable. I developed a theoretical framework for analyzing the information gain in infinite space using squashing functions.

Teaching Assistant

List of classes TA’ed
  • CS 60: Principles of Computer Science [Spring 2016]
  • CS 81: Computability and Logic [Spring 2016]
  • CS 158: Machine Learning [Fall 2017, Spring 2018, Spring 2019]
  • ENGR 155: Microprocessor Systems: Design and Application [Fall 2018]
  • ENGR 156: Introduction to Communication and Information Theory [Spring 2019]
  • MATH 189R: Mathematics of Big Data [Fall 2016]


List of classes taken
Computer Science
  • ENGR 156: Introduction to Communication and Information Theory [Spring 2018 by Prof. Timothy Tsai]
  • MATH 65: Differential Equations / Linear Algebra II [Summer 2016 by Prof. Dagan Karp]
  • MATH 189S: Parallel/High-Performance Computing [Fall 2016 by Prof. Jeho Park]
Physics, Chemistry, and Biology
  • JAPN 11: Japanese Conversation, Intermediate [Fall 2017]
Last updated: Jan 15, 2022