Jun-Kun Wang

I am a fourth year PhD student in CS at Georgia Tech. My advisor is Professor Jacob Abernethy. Before that, I was a research assistant at CSIE Department, National Taiwan University (NTU), where I worked with Prof. Shou-De Lin and also worked closely with Dr. Chi-Jen Lu at Academia Sinica. My PhD research focus on theoretical understanding of modern algorithms/techniques in optimization and deep learning (e.g. momentum, over-parametrization, accelerated gradient methods), as well as designing new algorithms with provable guarantees. I hope to study reinforcement learning in the future!

Fun fact 1: I like to run and enjoy long-distance running (especially along a river or bay). I run a lot!

Fun fact 2: I am a top x% reviewer of NeurIPS 2018, 2019. (x is the number that guarantees a registration of NeurIPS). Give me a paper, I can quickly provide a quite good review!

Reviewer of NeurIPS 2016,2017,2018,2019,2020, of ICML 2017,2018,2019,2020, of COLT 2017,2018,2019,2020, of ALT 2017,2018,2019,2020.

Publications: *Corresponding Author/*Presenting Author

Escape Saddle Points Faster with Stochastic Momentum.
*Jun-Kun Wang, Chi-Heng Lin, and Jacob Abernethy.
In International Conference on Learning Representations (ICLR) 2020.

Online Linear Optimization with Sparsity Constraints
*Jun-Kun Wang, * Chi-Jen Lu, and Shou-De Lin.
In International Conference on Algorithmic Learning Theory (ALT) 30, 2019.

Revisiting Projection-Free Optimization For Strongly Convex Constraint Sets
Jarrid Rector-Brooks, Jun-Kun Wang, and Barzan Mozafari.
In (AAAI) 33, 2019.

Acceleration through Optimistic No-Regret Dynamics
*Jun-Kun Wang and Jacob Abernethy.
In Annual Conference on Neural Information Processing Systems (NeurIPS) 2018 (Spotlight) paper

Faster Rates for Convex-Concave Games
(name order) Jacob Abernethy, Kevin Lai, Kfir Levy, and *Jun-Kun Wang.
In Computational Learning Theory (COLT) 2018.

On Frank-Wolfe and Equilibrium Computation
Jacob Abernethy and *Jun-Kun Wang.
In Annual Conference on Neural Information Processing Systems (NeurIPS) 2017 (Spotlight). paper supplementary

Efficient Sampling-based ADMM for Distributed Data
*Jun-Kun Wang, Shou-De Lin.
In IEEE International Conference on Data Science and Advanced Analytics (DSAA) 3, 2016. code

Parallel Least-Squares Policy Iteration
*Jun-Kun Wang, Shou-De Lin.
In IEEE International Conference on Data Science and Advanced Analytics (DSAA) 3, 2016.

Robust Inverse Covariance Estimation under Noisy Measurements
*Jun-Kun Wang, Shou-De Lin.
In International Conference on Machine Learning (ICML) 31, 2014. paper slide

Techical Reports:

1 Fast Logistic Bandit Jun-Kun Wang, Chi-Jen Lu, and Shou-De Lin paper (contains an error) Review