Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Christos Thrampoulidis

Christos Thrampoulidis

Label-Imbalanced and Group-Sensitive Classification under Overparameterization


Mar 02, 2021
Ganesh Ramachandra Kini, Orestis Paraskevas, Samet Oymak, Christos Thrampoulidis


  Access Paper or Ask Questions

Provable Benefits of Overparameterization in Model Compression: From Double Descent to Pruning Neural Networks


Dec 16, 2020
Xiangyu Chang, Yingcong Li, Samet Oymak, Christos Thrampoulidis

Figure 1 for Provable Benefits of Overparameterization in Model Compression: From Double Descent to Pruning Neural Networks
Figure 2 for Provable Benefits of Overparameterization in Model Compression: From Double Descent to Pruning Neural Networks
Figure 3 for Provable Benefits of Overparameterization in Model Compression: From Double Descent to Pruning Neural Networks
Figure 4 for Provable Benefits of Overparameterization in Model Compression: From Double Descent to Pruning Neural Networks
* to appear at AAAI 2021 
  Access Paper or Ask Questions

Decentralized Multi-Agent Linear Bandits with Safety Constraints


Dec 01, 2020
Sanae Amani, Christos Thrampoulidis

Figure 1 for Decentralized Multi-Agent Linear Bandits with Safety Constraints
Figure 2 for Decentralized Multi-Agent Linear Bandits with Safety Constraints
Figure 3 for Decentralized Multi-Agent Linear Bandits with Safety Constraints
Figure 4 for Decentralized Multi-Agent Linear Bandits with Safety Constraints

  Access Paper or Ask Questions

Benign Overfitting in Binary Classification of Gaussian Mixtures


Nov 18, 2020
Ke Wang, Christos Thrampoulidis

Figure 1 for Benign Overfitting in Binary Classification of Gaussian Mixtures
Figure 2 for Benign Overfitting in Binary Classification of Gaussian Mixtures
Figure 3 for Benign Overfitting in Binary Classification of Gaussian Mixtures
* 11 pages, 3 figures 
  Access Paper or Ask Questions

Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View


Nov 16, 2020
Christos Thrampoulidis, Samet Oymak, Mahdi Soltanolkotabi

Figure 1 for Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View
Figure 2 for Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View
Figure 3 for Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View
Figure 4 for Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View
* To Appear at NeurIPS 2020. 62 pages, 7 figures 
  Access Paper or Ask Questions

Asymptotic Behavior of Adversarial Training in Binary Classification


Oct 26, 2020
Hossein Taheri, Ramtin Pedarsani, Christos Thrampoulidis

Figure 1 for Asymptotic Behavior of Adversarial Training in Binary Classification
Figure 2 for Asymptotic Behavior of Adversarial Training in Binary Classification

  Access Paper or Ask Questions

Stage-wise Conservative Linear Bandits


Sep 30, 2020
Ahmadreza Moradipari, Christos Thrampoulidis, Mahnoosh Alizadeh

Figure 1 for Stage-wise Conservative Linear Bandits
Figure 2 for Stage-wise Conservative Linear Bandits
Figure 3 for Stage-wise Conservative Linear Bandits
* Thirty-fourth Conference on Neural Information Processing Systems, NeurIPS 2020 
 * 28 pages, 5 figures 
  Access Paper or Ask Questions

Optimal Combination of Linear and Spectral Estimators for Generalized Linear Models


Aug 11, 2020
Marco Mondelli, Christos Thrampoulidis, Ramji Venkataramanan

Figure 1 for Optimal Combination of Linear and Spectral Estimators for Generalized Linear Models
Figure 2 for Optimal Combination of Linear and Spectral Estimators for Generalized Linear Models
Figure 3 for Optimal Combination of Linear and Spectral Estimators for Generalized Linear Models
Figure 4 for Optimal Combination of Linear and Spectral Estimators for Generalized Linear Models
* 45 pages, 5 figures 
  Access Paper or Ask Questions

Fundamental Limits of Ridge-Regularized Empirical Risk Minimization in High Dimensions


Jul 05, 2020
Hossein Taheri, Ramtin Pedarsani, Christos Thrampoulidis

Figure 1 for Fundamental Limits of Ridge-Regularized Empirical Risk Minimization in High Dimensions
Figure 2 for Fundamental Limits of Ridge-Regularized Empirical Risk Minimization in High Dimensions
Figure 3 for Fundamental Limits of Ridge-Regularized Empirical Risk Minimization in High Dimensions
Figure 4 for Fundamental Limits of Ridge-Regularized Empirical Risk Minimization in High Dimensions

  Access Paper or Ask Questions

Exploring Weight Importance and Hessian Bias in Model Pruning


Jun 19, 2020
Mingchen Li, Yahya Sattar, Christos Thrampoulidis, Samet Oymak

Figure 1 for Exploring Weight Importance and Hessian Bias in Model Pruning
Figure 2 for Exploring Weight Importance and Hessian Bias in Model Pruning
Figure 3 for Exploring Weight Importance and Hessian Bias in Model Pruning
Figure 4 for Exploring Weight Importance and Hessian Bias in Model Pruning
* 28 pages 
  Access Paper or Ask Questions

Regret Bounds for Safe Gaussian Process Bandit Optimization


May 05, 2020
Sanae Amani, Mahnoosh Alizadeh, Christos Thrampoulidis

Figure 1 for Regret Bounds for Safe Gaussian Process Bandit Optimization
Figure 2 for Regret Bounds for Safe Gaussian Process Bandit Optimization
Figure 3 for Regret Bounds for Safe Gaussian Process Bandit Optimization
Figure 4 for Regret Bounds for Safe Gaussian Process Bandit Optimization
* 22 pages, 17 figures 
  Access Paper or Ask Questions

Sharp Asymptotics and Optimal Performance for Inference in Binary Models


Feb 26, 2020
Hossein Taheri, Ramtin Pedarsani, Christos Thrampoulidis

Figure 1 for Sharp Asymptotics and Optimal Performance for Inference in Binary Models
Figure 2 for Sharp Asymptotics and Optimal Performance for Inference in Binary Models
Figure 3 for Sharp Asymptotics and Optimal Performance for Inference in Binary Models
Figure 4 for Sharp Asymptotics and Optimal Performance for Inference in Binary Models

  Access Paper or Ask Questions

Analytic Study of Double Descent in Binary Classification: The Impact of Loss


Jan 30, 2020
Ganesh Kini, Christos Thrampoulidis

Figure 1 for Analytic Study of Double Descent in Binary Classification: The Impact of Loss
Figure 2 for Analytic Study of Double Descent in Binary Classification: The Impact of Loss
Figure 3 for Analytic Study of Double Descent in Binary Classification: The Impact of Loss
Figure 4 for Analytic Study of Double Descent in Binary Classification: The Impact of Loss

  Access Paper or Ask Questions

A Model of Double Descent for High-dimensional Binary Linear Classification


Nov 13, 2019
Zeyu Deng, Abla Kammoun, Christos Thrampoulidis

Figure 1 for A Model of Double Descent for High-dimensional Binary Linear Classification
Figure 2 for A Model of Double Descent for High-dimensional Binary Linear Classification
Figure 3 for A Model of Double Descent for High-dimensional Binary Linear Classification
Figure 4 for A Model of Double Descent for High-dimensional Binary Linear Classification
* Short version submitted to ICASSP 2020 
  Access Paper or Ask Questions

Safe Linear Thompson Sampling


Nov 06, 2019
Ahmadreza Moradipari, Sanae Amani, Mahnoosh Alizadeh, Christos Thrampoulidis

Figure 1 for Safe Linear Thompson Sampling
Figure 2 for Safe Linear Thompson Sampling
Figure 3 for Safe Linear Thompson Sampling
Figure 4 for Safe Linear Thompson Sampling

  Access Paper or Ask Questions

Linear Stochastic Bandits Under Safety Constraints


Aug 16, 2019
Sanae Amani, Mahnoosh Alizadeh, Christos Thrampoulidis

Figure 1 for Linear Stochastic Bandits Under Safety Constraints
Figure 2 for Linear Stochastic Bandits Under Safety Constraints
Figure 3 for Linear Stochastic Bandits Under Safety Constraints
* 23 pages, 7 figures 
  Access Paper or Ask Questions

Sharp Guarantees for Solving Random Equations with One-Bit Information


Aug 12, 2019
Hossein Taheri, Ramtin Pedarsani, Christos Thrampoulidis

Figure 1 for Sharp Guarantees for Solving Random Equations with One-Bit Information
Figure 2 for Sharp Guarantees for Solving Random Equations with One-Bit Information
Figure 3 for Sharp Guarantees for Solving Random Equations with One-Bit Information
Figure 4 for Sharp Guarantees for Solving Random Equations with One-Bit Information
* 12 pages, 4 figures 
  Access Paper or Ask Questions

Lifting high-dimensional nonlinear models with Gaussian regressors


Oct 26, 2018
Christos Thrampoulidis, Ankit Singh Rawat

Figure 1 for Lifting high-dimensional nonlinear models with Gaussian regressors
Figure 2 for Lifting high-dimensional nonlinear models with Gaussian regressors
* Improved the algorithm and expanded on its motivation; added simulation results 
  Access Paper or Ask Questions

The Squared-Error of Generalized LASSO: A Precise Analysis


Nov 06, 2013
Samet Oymak, Christos Thrampoulidis, Babak Hassibi

Figure 1 for The Squared-Error of Generalized LASSO: A Precise Analysis
Figure 2 for The Squared-Error of Generalized LASSO: A Precise Analysis
Figure 3 for The Squared-Error of Generalized LASSO: A Precise Analysis
Figure 4 for The Squared-Error of Generalized LASSO: A Precise Analysis

  Access Paper or Ask Questions