Alert button
Picture for Shahab Asoodeh

Shahab Asoodeh

Alert button

Sample-Optimal Locally Private Hypothesis Selection and the Provable Benefits of Interactivity

Dec 09, 2023
Alireza F. Pour, Hassan Ashtiani, Shahab Asoodeh

Viaarxiv icon

Privacy Loss of Noisy Stochastic Gradient Descent Might Converge Even for Non-Convex Losses

May 17, 2023
Shahab Asoodeh, Mario Diaz

Figure 1 for Privacy Loss of Noisy Stochastic Gradient Descent Might Converge Even for Non-Convex Losses
Figure 2 for Privacy Loss of Noisy Stochastic Gradient Descent Might Converge Even for Non-Convex Losses
Figure 3 for Privacy Loss of Noisy Stochastic Gradient Descent Might Converge Even for Non-Convex Losses
Viaarxiv icon

Contraction of Locally Differentially Private Mechanisms

Oct 24, 2022
Shahab Asoodeh, Huanyu Zhang

Figure 1 for Contraction of Locally Differentially Private Mechanisms
Viaarxiv icon

The Saddle-Point Accountant for Differential Privacy

Aug 20, 2022
Wael Alghamdi, Shahab Asoodeh, Flavio P. Calmon, Juan Felipe Gomez, Oliver Kosut, Lalitha Sankar, Fei Wei

Figure 1 for The Saddle-Point Accountant for Differential Privacy
Figure 2 for The Saddle-Point Accountant for Differential Privacy
Figure 3 for The Saddle-Point Accountant for Differential Privacy
Figure 4 for The Saddle-Point Accountant for Differential Privacy
Viaarxiv icon

Cactus Mechanisms: Optimal Differential Privacy Mechanisms in the Large-Composition Regime

Jun 25, 2022
Wael Alghamdi, Shahab Asoodeh, Flavio P. Calmon, Oliver Kosut, Lalitha Sankar, Fei Wei

Figure 1 for Cactus Mechanisms: Optimal Differential Privacy Mechanisms in the Large-Composition Regime
Figure 2 for Cactus Mechanisms: Optimal Differential Privacy Mechanisms in the Large-Composition Regime
Figure 3 for Cactus Mechanisms: Optimal Differential Privacy Mechanisms in the Large-Composition Regime
Figure 4 for Cactus Mechanisms: Optimal Differential Privacy Mechanisms in the Large-Composition Regime
Viaarxiv icon

Beyond Adult and COMPAS: Fairness in Multi-Class Prediction

Jun 15, 2022
Wael Alghamdi, Hsiang Hsu, Haewon Jeong, Hao Wang, P. Winston Michalak, Shahab Asoodeh, Flavio P. Calmon

Figure 1 for Beyond Adult and COMPAS: Fairness in Multi-Class Prediction
Figure 2 for Beyond Adult and COMPAS: Fairness in Multi-Class Prediction
Figure 3 for Beyond Adult and COMPAS: Fairness in Multi-Class Prediction
Figure 4 for Beyond Adult and COMPAS: Fairness in Multi-Class Prediction
Viaarxiv icon

Local Differential Privacy Is Equivalent to Contraction of $E_γ$-Divergence

Feb 02, 2021
Shahab Asoodeh, Maryam Aliakbarpour, Flavio P. Calmon

Figure 1 for Local Differential Privacy Is Equivalent to Contraction of $E_γ$-Divergence
Viaarxiv icon

Privacy Analysis of Online Learning Algorithms via Contraction Coefficients

Dec 20, 2020
Shahab Asoodeh, Mario Diaz, Flavio P. Calmon

Figure 1 for Privacy Analysis of Online Learning Algorithms via Contraction Coefficients
Figure 2 for Privacy Analysis of Online Learning Algorithms via Contraction Coefficients
Viaarxiv icon

Bottleneck Problems: Information and Estimation-Theoretic View

Nov 12, 2020
Shahab Asoodeh, Flavio Calmon

Figure 1 for Bottleneck Problems: Information and Estimation-Theoretic View
Figure 2 for Bottleneck Problems: Information and Estimation-Theoretic View
Figure 3 for Bottleneck Problems: Information and Estimation-Theoretic View
Figure 4 for Bottleneck Problems: Information and Estimation-Theoretic View
Viaarxiv icon

Three Variants of Differential Privacy: Lossless Conversion and Applications

Aug 14, 2020
Shahab Asoodeh, Jiachun Liao, Flavio P. Calmon, Oliver Kosut, Lalitha Sankar

Figure 1 for Three Variants of Differential Privacy: Lossless Conversion and Applications
Figure 2 for Three Variants of Differential Privacy: Lossless Conversion and Applications
Figure 3 for Three Variants of Differential Privacy: Lossless Conversion and Applications
Figure 4 for Three Variants of Differential Privacy: Lossless Conversion and Applications
Viaarxiv icon