Chaobing Song (宋朝兵)

I am a postdoc researcher at University of Wisconsin-Madison with Prof. Jelena Diakonikolas and Prof. Stephen J Wright. Now I am visiting Simons Institute for the Theory of Computing, Berkeley. My research interests include optimization and machine learning. I obtained my Ph.D. degree at Tsinghua University in 2020 supervised by Prof. Yi Ma from University of California, Berkeley (who is an affiliate professor at Tsinghua-Berkeley Shenzhen Institute, Tsinghua University). I spent two wonderful years at Berkeley and finished my Ph.D. thesis there. The main focus of my current research is to propose and apply a general, powerful and concise framework [arXiv] for optimization problems in machine learning. The development based on this framework substantially improves complexity results in many well-known settings.

Publications

Preprints

• Chaobing Song, Cheuk Yin Lin, Stephen J. Wright, Jelena Diakonikolas. Coordinate Linear Variance Reduction for Generalized Linear Programming. arXiv preprint, arXiv:2111.01842, 2021. [arXiv]

• Chaobing Song, Jelena Diakonikolas. Fast cyclic coordinate dual averaging with extrapolation for generalized variational inequalities. arXiv preprint, arXiv:2102.13244, 2021. [arXiv]

• Chaobing Song, Yong Jiang, Yi Ma. Breaking the O(1/\epsilon) optimal rate for a class of minimax optimization. arXiv preprint arXiv:2003.11758, 2020. [arXiv]

Conferences

• Chaobing Song, Stephen J Wright, Jelena Diakonikolas. Variance reduction via primal-dual accelerated dual averaging for nonsmooth convex finite-sums. In Proceedings of the 38th International Conference on Machine Learning (ICML), 2021. (Long Presentation, Acceptance Rate: 3.0%) [arXiv] [slides]

• Chaobing Song, Yong Jiang, Yi Ma. Variance reduction via accelerated dual averaging for finite-sum optimization. In Proceedings of the 34th International Conference on Neural Information Processing Systems (NeurIPS), 2020. [arXiv][slides]

• Chaobing Song, Zhengyuan Zhou, Yichao Zhou, Yong Jiang, Yi Ma. Optimistic dual extrapolation for coherent non-monotone variational inequalities. In Proceedings of the 34th International Conference on Neural Information Processing Systems (NeurIPS), 2020. [arXiv]

• Yaodong Yu, Kwan Ho Ryan Chan, Chong You, Chaobing Song, Yi Ma. Learning diverse and discriminative representations via the principle of maximal coding rate reduction. In Proceedings of the 34th International Conference on Neural Information Processing Systems (NeurIPS), 2020. [arXiv]

• Chaobing Song, Shaobo Cui, Yong Jiang, Shu-Tao Xia. Accelerated stochastic greedy coordinate descent by soft thresholding projection onto simplex. In Proceedings of the 31th International Conference on Neural Information Processing Systems (NeurIPS), pages 4838–4847, 2017 (Spotlight, Acceptance Rate: 3.4%). [arXiv]

• Chaobing Song, Shu-Tao Xia, Xin-Ji Liu. Subspace thresholding pursuit: A reconstruction algorithm for compressed sensing. IEEE International Symposium on Information Theory (ISIT), 2015. [arXiv]

• Tao Dai, Chaobing Song, Ji-Ping Zhang, Shu-Tao Xia. PMPA: A patch-based multiscale products algorithm for image denoising. IEEE International Conference on Image Processing (ICIP), 2015. [link]

Journals

• Chaobing Song, Yong Jiang, Yi Ma. Unified acceleration of high-order algorithms under H\"{o}lder continuity and uniform convexity. SIAM Journal on Optimization, 2021 (to appear). [arXiv]

• Chaobing Song, Shu-Tao Xia. Sparse signal recovery by \ell_q minimization under restricted isometry property. IEEE Signal Processing Letters, 2014, 21 (9). [arXiv]

• Chaobing Song, Shu-Tao Xia, Xin-Ji Liu. Improved analysis for subspace pursuit algorithm in terms of restricted isometry constant. IEEE Signal Processing Letters, 2014, 21 (11). [arXiv]

Conference Workshop

• Chaobing Song, Yong Jiang, Yi Ma. Newton dual extrapolation for non-monotone variational inequalities. OPT-ML Workshop of 37th International Conference of Machine Learning (ICML), 2020. [link]