no code implementations • 28 Jan 2024 • Hongjian Wang, Aaditya Ramdas
We present new concentration inequalities for either martingale dependent or exchangeable random symmetric matrices under a variety of tail conditions, encompassing now-standard Chernoff bounds to self-normalized heavy-tailed settings.
no code implementations • 14 Nov 2023 • Ben Chugg, Hongjian Wang, Aaditya Ramdas
We derive and study time-uniform confidence spheres -- confidence sphere sequences (CSSs) -- which contain the mean of random vectors with high probability simultaneously across all sample sizes.
no code implementations • 5 Oct 2023 • Hongjian Wang, Aaditya Ramdas
These are respectively obtained by swapping Lai's flat mixture for a Gaussian mixture, and swapping the right Haar mixture over $\sigma$ with the maximum likelihood estimate under the null, as done in universal inference.
no code implementations • 12 Sep 2023 • Chunqing Ruan, Hongjian Wang
Parameter efficient transfer learning (PETL) is an emerging research spot that aims to adapt large-scale pre-trained models to downstream tasks.
no code implementations • 3 Apr 2023 • Hongjian Wang, Aaditya Ramdas
Following the initial work by Robbins, we rigorously present an extended theory of nonnegative supermartingales, requiring neither integrability nor finiteness.
1 code implementation • CVPR 2023 • Yunfan Lu, Zipeng Wang, Minjie Liu, Hongjian Wang, Lin Wang
In addition, we collect a real-world dataset with spatially aligned events and RGB frames.
no code implementations • 7 Feb 2023 • Ben Chugg, Hongjian Wang, Aaditya Ramdas
We present a unified framework for deriving PAC-Bayesian generalization bounds.
no code implementations • 23 Jan 2023 • Hongjian Wang, Aaditya Ramdas
Confidence sequences are confidence intervals that can be sequentially tracked, and are valid at arbitrary data-dependent stopping times.
no code implementations • NeurIPS 2021 • Hongjian Wang, Mert Gürbüzbalaban, Lingjiong Zhu, Umut Şimşekli, Murat A. Erdogdu
In this paper, we provide convergence guarantees for SGD under a state-dependent and heavy-tailed noise with a potentially infinite variance, for a class of strongly convex objectives.
no code implementations • 31 Dec 2020 • Hongjian Wang, Qi Li, Lanbo Zhang, Yue Lu, Steven Yoo, Srinivas Vadrevu, Zhenhui Li
Historical features are important in ads click-through rate (CTR) prediction, because they account for past engagements between users and ads.
no code implementations • 29 Aug 2019 • Guanjie Zheng, Mengqi Liu, Tao Wen, Hongjian Wang, Huaxiu Yao, Susan L. Brantley, Zhenhui Li
In the face of growing needs for water and energy, a fundamental understanding of the environmental impacts of human activities becomes critical for managing water and energy resources, remedying water pollution, and making regulatory policy wisely.
no code implementations • 28 Dec 2015 • Hongjian Wang, Zhenhui Li, Yu-Hsuan Kuo, Dan Kifer
The increased availability of large-scale trajectory data around the world provides rich information for the study of urban dynamics.