Junwen Qiu

Interested in nonconvex stochastic optimization.

[Junwen Qiu @JunwenQiu]

@michaelchchoi Happy birthday ~

[Michael Choi @michaelchchoi]

One year older, one year wiser.
#MichaelChoi
#MarkovChains

[arXiv math.OC Optimization and Control @mathOCb]

Junwen Qiu, Ziyang Zeng, Leilei Mei, Junyu Zhang: A New Kernel Regularity Condition for Distributed Mirror Descent: Broader Coverage and Simpler Analysis https://t.co/vHEiGAkeJq https://t.co/p0ik1vj2RV https://t.co/cbXVfD8wBF

[Stat.ML Papers @StatMLPapers]

Shuffling the Stochastic Mirror Descent via Dual Lipschitz Continuity and Kernel Conditioning https://t.co/h4cLPjBqd2

[Junwen Qiu @JunwenQiu]

@michaelchchoi @StatMLPapers Thanks~

[Lexiao Lai @lailexiao]

Do iterates of the subgradient method converge for nonconvex, nonsmooth objectives?

Yes, for tame/definable functions, covering most ML/AI training objectives. Check out our new preprint https://t.co/gFVKBVAQ2E (w. my PhD student @MingzhiS13149) #nonconvex #optimization

1/3

[Journal of Machine Learning Research @JmlrOrg]

'A New Random Reshuffling Method for Nonsmooth Nonconvex Finite-sum Optimization', by Junwen Qiu, Xiao Li, Andre Milzarek.

https://t.co/Mt4BSyuFQG

#complexity #iterations #nonconvex

[Michael Choi @michaelchchoi]

Further updated today:

Title: Geometry and factorization of multivariate Markov chains with applications to MCMC acceleration

arXiv: https://t.co/mv2WOIVHwp

[Michael Choi @michaelchchoi]

[Group-averaged Markov chains] Updated, with a more complete bibliography.

arXiv: https://t.co/CYuaB3ssDm

[Junwen Qiu @JunwenQiu]

@michaelchchoi Looks very fancy!

[XY Han @XYHan_]

👏👏👏 Much needed and overdue.
A huge personal pain point for me as an opt researcher is that popular constrained opt solvers (cvxpy, gurobi, mosek, etc) require specialized syntax for the constraints and end up moving back to CPU and so can't advantage of GPU matmult... (1/2)

[Tony S.F. @tonysilveti]

I don't think this book is so well-known since it's only recently been made available but it's got very nice material and illustrations, especially in the set-valued analysis section: https://t.co/rdsPzvoHUq

[arXiv math.OC Optimization and Control @mathOCb]

Junwen Qiu, Bohao Ma, Xiao Li, Andre Milzarek: A KL-based Analysis Framework with Applications to Non-Descent Optimization Methods https://t.co/tiMLJvKHoq https://t.co/J19iIbi8Ld

[Junwen Qiu @JunwenQiu]

@michaelchchoi @mathOCb Yeah! Andre’s group is very productive recently!

[Junwen Qiu @JunwenQiu]

@michaelchchoi @mathOCb Thanks!

[Junwen Qiu @JunwenQiu]

@damekdavis Congrats! Also, Liwei is a very strong climber!

[arXiv math.OC Optimization and Control @mathOCb]

Junwen Qiu, Andre Milzarek: Random Reshuffling with Momentum for Nonconvex Problems: Iteration Complexity and Last Iterate Convergence https://t.co/0laNxChE1L https://t.co/nmyqZRp1bY

[Junwen Qiu @JunwenQiu]

@michaelchchoi @mathOCb Thank you! Michael

[Gabriel Peyré @gabrielpeyre]

Gradient descent does not always converge. Trajectories can have infinite length. Picture & insights from @GuillaumeG_ https://t.co/KMH3CCpvCM

[Oded Rechavi @OdedRechavi]

Running into your professor outside the university