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 Mehrdad Mahdavi

Learn Locally, Correct Globally: A Distributed Algorithm for Training Graph Neural Networks


Dec 07, 2021
Morteza Ramezani, Weilin Cong, Mehrdad Mahdavi, Mahmut T. Kandemir, Anand Sivasubramaniam


  Access Paper or Ask Questions

Dynamic Graph Representation Learning via Graph Transformer Networks


Nov 19, 2021
Weilin Cong, Yanhong Wu, Yuandong Tian, Mengting Gu, Yinglong Xia, Mehrdad Mahdavi, Chun-cheng Jason Chen


  Access Paper or Ask Questions

On Provable Benefits of Depth in Training Graph Convolutional Networks


Oct 28, 2021
Weilin Cong, Morteza Ramezani, Mehrdad Mahdavi


  Access Paper or Ask Questions

Meta-learning with an Adaptive Task Scheduler


Oct 26, 2021
Huaxiu Yao, Yu Wang, Ying Wei, Peilin Zhao, Mehrdad Mahdavi, Defu Lian, Chelsea Finn

* Accepted by NeurIPS 2021 

  Access Paper or Ask Questions

Local SGD Optimizes Overparameterized Neural Networks in Polynomial Time


Jul 22, 2021
Yuyang Deng, Mehrdad Mahdavi


  Access Paper or Ask Questions

Pareto Efficient Fairness in Supervised Learning: From Extraction to Tracing


Apr 04, 2021
Mohammad Mahdi Kamani, Rana Forsati, James Z. Wang, Mehrdad Mahdavi


  Access Paper or Ask Questions

On the Importance of Sampling in Learning Graph Convolutional Networks


Mar 03, 2021
Weilin Cong, Morteza Ramezani, Mehrdad Mahdavi


  Access Paper or Ask Questions

Local Stochastic Gradient Descent Ascent: Convergence Analysis and Communication Efficiency


Feb 25, 2021
Yuyang Deng, Mehrdad Mahdavi

* This paper has been accepted to AISTATS 2021 

  Access Paper or Ask Questions

Distributionally Robust Federated Averaging


Feb 25, 2021
Yuyang Deng, Mohammad Mahdi Kamani, Mehrdad Mahdavi

* Advances in Neural Information Processing Systems (NeurIPS), Vol. 33, 2020 
* Published in NeurIPS 2020: https://proceedings.neurips.cc/paper/2020/hash/ac450d10e166657ec8f93a1b65ca1b14-Abstract.html 

  Access Paper or Ask Questions

Communication-efficient k-Means for Edge-based Machine Learning


Feb 08, 2021
Hanlin Lu, Ting He, Shiqiang Wang, Changchang Liu, Mehrdad Mahdavi, Vijaykrishnan Narayanan, Kevin S. Chan, Stephen Pasteris


  Access Paper or Ask Questions

Online Structured Meta-learning


Oct 22, 2020
Huaxiu Yao, Yingbo Zhou, Mehrdad Mahdavi, Zhenhui Li, Richard Socher, Caiming Xiong

* Accepted by NeurIPS 2020 

  Access Paper or Ask Questions

Federated Learning with Compression: Unified Analysis and Sharp Guarantees


Jul 02, 2020
Farzin Haddadpour, Mohammad Mahdi Kamani, Aryan Mokhtari, Mehrdad Mahdavi

* 52 pages 

  Access Paper or Ask Questions

Minimal Variance Sampling with Provable Guarantees for Fast Training of Graph Neural Networks


Jun 24, 2020
Weilin Cong, Rana Forsati, Mahmut Kandemir, Mehrdad Mahdavi


  Access Paper or Ask Questions

Adaptive Personalized Federated Learning


Mar 30, 2020
Yuyang Deng, Mohammad Mahdi Kamani, Mehrdad Mahdavi


  Access Paper or Ask Questions

On the Convergence of Local Descent Methods in Federated Learning


Dec 06, 2019
Farzin Haddadpour, Mehrdad Mahdavi

* 47 pages, "Updates from v1: A technical error in Lemma B3 is corrected" 

  Access Paper or Ask Questions

Efficient Fair Principal Component Analysis


Nov 12, 2019
Mohammad Mahdi Kamani, Farzin Haddadpour, Rana Forsati, Mehrdad Mahdavi


  Access Paper or Ask Questions

Local SGD with Periodic Averaging: Tighter Analysis and Adaptive Synchronization


Oct 30, 2019
Farzin Haddadpour, Mohammad Mahdi Kamani, Mehrdad Mahdavi, Viveck R. Cadambe

* Paper accepted to NeurIPS 2019 

  Access Paper or Ask Questions

Learning Feature Nonlinearities with Non-Convex Regularized Binned Regression


May 20, 2017
Samet Oymak, Mehrdad Mahdavi, Jiasi Chen

* 22 pages, 7 figures 

  Access Paper or Ask Questions

Sketching Meets Random Projection in the Dual: A Provable Recovery Algorithm for Big and High-dimensional Data


Oct 10, 2016
Jialei Wang, Jason D. Lee, Mehrdad Mahdavi, Mladen Kolar, Nathan Srebro


  Access Paper or Ask Questions

Train and Test Tightness of LP Relaxations in Structured Prediction


Apr 27, 2016
Ofer Meshi, Mehrdad Mahdavi, Adrian Weller, David Sontag

* To appear in ICML 2016 

  Access Paper or Ask Questions

Matrix Factorization with Explicit Trust and Distrust Relationships


Aug 02, 2014
Rana Forsati, Mehrdad Mahdavi, Mehrnoush Shamsfard, Mohamed Sarwat

* ACM Transactions on Information Systems 

  Access Paper or Ask Questions

Exploiting Smoothness in Statistical Learning, Sequential Prediction, and Stochastic Optimization


Jul 19, 2014
Mehrdad Mahdavi

* Ph.D. Thesis 

  Access Paper or Ask Questions

Recovering the Optimal Solution by Dual Random Projection


Feb 21, 2014
Lijun Zhang, Mehrdad Mahdavi, Rong Jin, Tianbao Yang, Shenghuo Zhu

* The 26th Annual Conference on Learning Theory (COLT 2013) 

  Access Paper or Ask Questions

Excess Risk Bounds for Exponentially Concave Losses


Feb 08, 2014
Mehrdad Mahdavi, Rong Jin


  Access Paper or Ask Questions

Binary Excess Risk for Smooth Convex Surrogates


Feb 07, 2014
Mehrdad Mahdavi, Lijun Zhang, Rong Jin


  Access Paper or Ask Questions

Beating the Minimax Rate of Active Learning with Prior Knowledge


Feb 06, 2014
Lijun Zhang, Mehrdad Mahdavi, Rong Jin


  Access Paper or Ask Questions

MixedGrad: An O(1/T) Convergence Rate Algorithm for Stochastic Smooth Optimization


Jul 26, 2013
Mehrdad Mahdavi, Rong Jin


  Access Paper or Ask Questions

An Efficient Primal-Dual Prox Method for Non-Smooth Optimization


Jul 26, 2013
Tianbao Yang, Mehrdad Mahdavi, Rong Jin, Shenghuo Zhu


  Access Paper or Ask Questions