Alert button
Picture for Ashok Cutkosky

Ashok Cutkosky

Alert button

When, Why and How Much? Adaptive Learning Rate Scheduling by Refinement

Oct 11, 2023
Aaron Defazio, Ashok Cutkosky, Harsh Mehta, Konstantin Mishchenko

Figure 1 for When, Why and How Much? Adaptive Learning Rate Scheduling by Refinement
Figure 2 for When, Why and How Much? Adaptive Learning Rate Scheduling by Refinement
Figure 3 for When, Why and How Much? Adaptive Learning Rate Scheduling by Refinement
Figure 4 for When, Why and How Much? Adaptive Learning Rate Scheduling by Refinement
Viaarxiv icon

Improving Adaptive Online Learning Using Refined Discretization

Sep 27, 2023
Zhiyu Zhang, Heng Yang, Ashok Cutkosky, Ioannis Ch. Paschalidis

Viaarxiv icon

Unconstrained Online Learning with Unbounded Losses

Jun 08, 2023
Andrew Jacobsen, Ashok Cutkosky

Viaarxiv icon

Mechanic: A Learning Rate Tuner

Jun 02, 2023
Ashok Cutkosky, Aaron Defazio, Harsh Mehta

Figure 1 for Mechanic: A Learning Rate Tuner
Figure 2 for Mechanic: A Learning Rate Tuner
Figure 3 for Mechanic: A Learning Rate Tuner
Figure 4 for Mechanic: A Learning Rate Tuner
Viaarxiv icon

Optimal Stochastic Non-smooth Non-convex Optimization through Online-to-Non-convex Conversion

Feb 11, 2023
Ashok Cutkosky, Harsh Mehta, Francesco Orabona

Viaarxiv icon

Unconstrained Dynamic Regret via Sparse Coding

Jan 31, 2023
Zhiyu Zhang, Ashok Cutkosky, Ioannis Ch. Paschalidis

Figure 1 for Unconstrained Dynamic Regret via Sparse Coding
Viaarxiv icon

Differentially Private Image Classification from Features

Nov 24, 2022
Harsh Mehta, Walid Krichene, Abhradeep Thakurta, Alexey Kurakin, Ashok Cutkosky

Figure 1 for Differentially Private Image Classification from Features
Figure 2 for Differentially Private Image Classification from Features
Figure 3 for Differentially Private Image Classification from Features
Figure 4 for Differentially Private Image Classification from Features
Viaarxiv icon

Parameter-free Regret in High Probability with Heavy Tails

Oct 25, 2022
Jiujia Zhang, Ashok Cutkosky

Figure 1 for Parameter-free Regret in High Probability with Heavy Tails
Viaarxiv icon

Differentially Private Online-to-Batch for Smooth Losses

Oct 12, 2022
Qinzi Zhang, Hoang Tran, Ashok Cutkosky

Viaarxiv icon