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

Add to Chrome

Add to Firefox

Picture for Barzan Mozafari

Barzan Mozafari

Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers

May 28, 2022
Rui Liu, Young Jin Kim, Alexandre Muzio, Barzan Mozafari, Hany Hassan Awadalla

Figure 1 for Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers

Figure 2 for Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers

Figure 3 for Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers

Figure 4 for Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers

* Accepted to ICML 2022

Access Paper or Ask Questions

Transformer with Memory Replay

May 19, 2022
Rui Liu, Barzan Mozafari

Figure 1 for Transformer with Memory Replay

Figure 2 for Transformer with Memory Replay

Figure 3 for Transformer with Memory Replay

Figure 4 for Transformer with Memory Replay

* Accepted to AAAI 2022

Access Paper or Ask Questions

Adam with Bandit Sampling for Deep Learning

Oct 24, 2020
Rui Liu, Tianyi Wu, Barzan Mozafari

Figure 1 for Adam with Bandit Sampling for Deep Learning

Figure 2 for Adam with Bandit Sampling for Deep Learning

Figure 3 for Adam with Bandit Sampling for Deep Learning

Figure 4 for Adam with Bandit Sampling for Deep Learning

* Accepted to NeurIPS 2020 as spotlight presentation

Access Paper or Ask Questions

BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees

Dec 26, 2018
Yongjoo Park, Jingyi Qing, Xiaoyang Shen, Barzan Mozafari

Figure 1 for BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees

Figure 2 for BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees

Figure 3 for BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees

Figure 4 for BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees

* 22 pages, SIGMOD 2019

Access Paper or Ask Questions

A Bandit Approach to Maximum Inner Product Search

Dec 15, 2018
Rui Liu, Tianyi Wu, Barzan Mozafari

Figure 1 for A Bandit Approach to Maximum Inner Product Search

Figure 2 for A Bandit Approach to Maximum Inner Product Search

Figure 3 for A Bandit Approach to Maximum Inner Product Search

Figure 4 for A Bandit Approach to Maximum Inner Product Search

* AAAI 2019

Access Paper or Ask Questions

Revisiting Projection-Free Optimization for Strongly Convex Constraint Sets

Nov 14, 2018
Jarrid Rector-Brooks, Jun-Kun Wang, Barzan Mozafari

Figure 1 for Revisiting Projection-Free Optimization for Strongly Convex Constraint Sets

Figure 2 for Revisiting Projection-Free Optimization for Strongly Convex Constraint Sets

Figure 3 for Revisiting Projection-Free Optimization for Strongly Convex Constraint Sets

Figure 4 for Revisiting Projection-Free Optimization for Strongly Convex Constraint Sets

* Appearing in AAAI-19, 27 pages

Access Paper or Ask Questions

Database Learning: Toward a Database that Becomes Smarter Every Time

Mar 28, 2017
Yongjoo Park, Ahmad Shahab Tajik, Michael Cafarella, Barzan Mozafari

Figure 1 for Database Learning: Toward a Database that Becomes Smarter Every Time

Figure 2 for Database Learning: Toward a Database that Becomes Smarter Every Time

Figure 3 for Database Learning: Toward a Database that Becomes Smarter Every Time

Figure 4 for Database Learning: Toward a Database that Becomes Smarter Every Time

* This manuscript is an extended report of the work published in ACM SIGMOD conference 2017

Access Paper or Ask Questions

Active Learning for Crowd-Sourced Databases

Dec 20, 2014
Barzan Mozafari, Purnamrita Sarkar, Michael J. Franklin, Michael I. Jordan, Samuel Madden

Figure 1 for Active Learning for Crowd-Sourced Databases

Figure 2 for Active Learning for Crowd-Sourced Databases

Figure 3 for Active Learning for Crowd-Sourced Databases

Figure 4 for Active Learning for Crowd-Sourced Databases

* A shorter version of this manuscript has been published in Proceedings of Very Large Data Bases 2015, entitled "Scaling Up Crowd-Sourcing to Very Large Datasets: A Case for Active Learning"

Access Paper or Ask Questions