An Optimal Reduction of TV-Denoising to Adaptive Online Learning
Baby, Dheeraj, Zhao, Xuandong, Wang, Yu-Xiang
We consider the problem of estimating a function from $n$ noisy samples whose discrete Total Variation (TV) is bounded by $C_n$. We reveal a deep connection to the seemingly disparate problem of Strongly Adaptive online learning (Daniely et al, 2015) and provide an $O(n \log n)$ time algorithm that attains the near minimax optimal rate of $\tilde O (n^{1/3}C_n^{2/3})$ under squared error loss. The resulting algorithm runs online and optimally adapts to the unknown smoothness parameter $C_n$. This leads to a new and more versatile alternative to wavelets-based methods for (1) adaptively estimating TV bounded functions; (2) online forecasting of TV bounded trends in time series.
Jan-26-2021
- Country:
- North America > United States
- New Mexico (0.04)
- New York (0.04)
- West Virginia (0.04)
- South Carolina (0.04)
- Alaska (0.04)
- Maine (0.04)
- Arkansas (0.04)
- Florida (0.04)
- Alabama (0.04)
- Minnesota (0.04)
- Utah (0.04)
- Nevada (0.04)
- Indiana (0.04)
- Texas (0.04)
- Missouri (0.04)
- Wisconsin (0.04)
- Arizona (0.04)
- South Dakota (0.04)
- Oregon (0.04)
- Connecticut (0.04)
- Maryland (0.04)
- Kansas (0.04)
- Hawaii (0.04)
- District of Columbia (0.04)
- North Carolina (0.04)
- Michigan (0.04)
- Tennessee (0.04)
- New Hampshire (0.04)
- North Dakota (0.04)
- Rhode Island (0.04)
- Iowa (0.04)
- Oklahoma (0.04)
- Ohio (0.04)
- New Jersey (0.04)
- Louisiana (0.04)
- Vermont (0.04)
- Nebraska (0.04)
- Idaho (0.04)
- Virginia (0.04)
- Illinois (0.04)
- Colorado (0.04)
- Pennsylvania (0.04)
- Wyoming (0.04)
- Kentucky (0.04)
- Mississippi (0.04)
- Massachusetts (0.04)
- California (0.04)
- Montana (0.04)
- Europe
- United Kingdom > England
- Cambridgeshire > Cambridge (0.04)
- Spain > Galicia
- Madrid (0.04)
- United Kingdom > England
- North America > United States
- Genre:
- Research Report (0.50)
- Industry:
- Technology: