HELLO, I'M

​Keyi Chen

WechatIMG18.jpeg

​Keyi Chen

Hi, I am currently third year PhD student in Boston University, Computer Science Department, supervised by Francesco Orabona​. Previously, I was at Stony Brook University.

​I'm interested in machine learning theory. Currently I focus on parameter free algorithms.

 

About

MY RESEARCH INTEREST

With the exponential growth of information, algorithms are trained on large amounts of data. Being efficient is an essential problem to be considered. Most of the time, you find algorithms discuss about optimal tuned result. However, tuning process is a headache all the time. I'm interested in designing parameter-free algorithms that do not require tuning learning rate, but could also achieve optimal empirical result. 

 

Education

Publications

2019 Spring - Now

Preprints

Boston University – Boston, MA; 

PhD; Computer Science

2016–2019

Stony Brook University – Stony Brook, NY

MS; Statistics

2011–2015

Shandong University, Shandong, China 

BS; Statistics

Better Parameter-Free Stochastic Optimization with ODE Updates for Coin Betting (arXiv)

Parameter-free stochastic gradient descent (PFSGD) algorithms do not require setting learning rates while achieving optimal theoretical performance. In practical applications, however, there remains an empirical gap between tuned stochastic gradient descent (SGD) and PFSGD. In this paper, we close the empirical gap with a new parameter-free algorithm based on continuous-time Coin-Betting on truncated models. The new update is derived through the solution of an Ordinary Differential Equation (ODE) and solved in a closed form. We show empirically that this new parameter-free algorithm outperforms algorithms with the ``best default'' learning rates and almost matches the performance of finely tuned baselines without anything to tune.

 

Work Experience

Optimal Lab, Boston, Research Assistant, Advisor: Francesco Orabona
  • Jan.2019 - Now

  • Create novel parameter-free method and prove its superiority over existing algorithms theoretically. 

  • Conduct extensive experiments on realistic datasets to demonstrate  our algorithm's merits 

Stony Brook University, Stony Brook, Teaching Assistant    
  • Jan.2018 - Jan.2019

  • Performed academic tutoring during office hours, responsible for grading assignments and exams

  • Assisted head faculty member with classroom instruction material, exams, and record keeping

Citibank, Beijing (CN), Academic Intern
  • Jun.2015 - Aug.2015

  • Studied fraud management, build various models to recognize fraudulent credit card transactions

  • Build customer models by data mining and help marketing professional to target marketing campaigns more accurately