Yilong's Homepage
Home
Yosemite, 2022

Welcome to my homepage! I’m a currently third year CS PhD student at Rice University, advised by Prof. Ankit Patel.

My research focuses on building and improving machine learning and deep learning models for science problems in an explainable way. The fields of subjects involve quantum many-body physics, computational chemical sensing, medicine, etc. To be more specific, how can we build models that do not require too much domain knowledge, but are able to learn and exploit various types of data in different fields of Science? And meanwhile, can we still have sufficient explainability of these models by replacing deep neural networks with shallower? Maybe we are able to achieve the same goal without neural networks, by learning from their success. I also worked on understanding GANs’ behavior in the function space in the rich and kernel regimes. I’m interested in adversarial learning schemes in general.

I received my Master’s degree from University of Pennsylvania, where I studied Data Science and worked with Prof. Shivani Agarwal on Learning the Tree Structure of a Nested Logit Choice Model from Data. I also worked for Prof. Barry Silverman on Addressing Disparities in Healthcare Access and Outcomes among Chronically Ill Older Adults: Assessing the Feasibility of an Agent-Based Modeling Approach (visualization of the geographic data used in the project). I received my Bachelor’s degree from Tsinghua University, where I studied Industrial Engineering.

Out of all my hobbies, I’m proudest of my efforts in music composition. I was a bassoon and saxophone player in Tsinghua University Symphonic Band and hence I had the opportunity to write pieces for different sizes of bands, from woodwind duets, saxophone quartets, woodwind quintets, to the full band. In my last year as an undergrad, I rearranged the La La Land Suite and it is performed as a formal piece on our annual concert (video). In my free time, I also play tennis, badminton, and go hiking. My favorite singer is Kotoha.

Publications

Yilong Ju, Jonathan Minoff, Shah Saad Alam, Fabio Anselmi, Han Pu, Ankit Patel. Looking Under the Hood: How Convolutional Neural Networks Successfully Approximate Quantum Spin Hamiltonians. In progress.

Mary Bajomo*, Yilong Ju*, Jingyi Zhou, Simina Elefterescu, Corbin Farr, Yiping Zhao, Oara Neumann, Peter Nordlander, Ankit Patel**, Naomi Halas**. Detection and Identification of Polycyclic Aromatic Hydrocarbons Using Surface Enhanced Raman Spectroscopy and Effective Characteristic Peak Extraction. In progress. *Co-first authors. **Co-last authors.

Jamie L.S. Waugh, Raajen Patel, Yilong Ju, Ankit Patel, Craig Rusin, Parag Jain. A Novel Automated Junctional Ectopic Tachycardia Detection Tool for Children with Congenital Heart Disease. Heart Rhythm O2 (2022).

Josue Ortega Caro, Yilong Ju, Ryan Pyle, Sourav Dey, Wieland Brendel, Fabio Anselmi, Ankit Patel. Local Convolutions Cause an Implicit Bias towards High Frequency Adversarial Examples. arXiv preprint

Yilong Ju, Weili Nie, Vikram Maheshri, Ankit Patel. Implicit Bias towards the Kernel Regime Causes Mode Collapse in GANs. Under Review.

Yilong Ju. Learning the Tree Structure of a Nested Logit Choice Model from Data. University of Pennsylvania Master's Thesis