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

Add to Chrome

Add to Firefox

Picture for Prateek Mittal

Prateek Mittal

A Light Recipe to Train Robust Vision Transformers

Sep 15, 2022
Edoardo Debenedetti, Vikash Sehwag, Prateek Mittal

Figure 1 for A Light Recipe to Train Robust Vision Transformers

Figure 2 for A Light Recipe to Train Robust Vision Transformers

Figure 3 for A Light Recipe to Train Robust Vision Transformers

Figure 4 for A Light Recipe to Train Robust Vision Transformers

* Code available at https://github.com/dedeswim/vits-robustness-torch

Access Paper or Ask Questions

Understanding Robust Learning through the Lens of Representation Similarities

Jun 20, 2022
Christian Cianfarani, Arjun Nitin Bhagoji, Vikash Sehwag, Ben Zhao, Prateek Mittal

Figure 1 for Understanding Robust Learning through the Lens of Representation Similarities

Figure 2 for Understanding Robust Learning through the Lens of Representation Similarities

Figure 3 for Understanding Robust Learning through the Lens of Representation Similarities

Figure 4 for Understanding Robust Learning through the Lens of Representation Similarities

* 35 pages, 29 figures

Access Paper or Ask Questions

Neurotoxin: Durable Backdoors in Federated Learning

Jun 12, 2022
Zhengming Zhang, Ashwinee Panda, Linyue Song, Yaoqing Yang, Michael W. Mahoney, Joseph E. Gonzalez, Kannan Ramchandran, Prateek Mittal

Figure 1 for Neurotoxin: Durable Backdoors in Federated Learning

Figure 2 for Neurotoxin: Durable Backdoors in Federated Learning

Figure 3 for Neurotoxin: Durable Backdoors in Federated Learning

Figure 4 for Neurotoxin: Durable Backdoors in Federated Learning

* Appears in ICML 2022

Access Paper or Ask Questions

Fight Poison with Poison: Detecting Backdoor Poison Samples via Decoupling Benign Correlations

May 26, 2022
Xiangyu Qi, Tinghao Xie, Saeed Mahloujifar, Prateek Mittal

Figure 1 for Fight Poison with Poison: Detecting Backdoor Poison Samples via Decoupling Benign Correlations

Figure 2 for Fight Poison with Poison: Detecting Backdoor Poison Samples via Decoupling Benign Correlations

Figure 3 for Fight Poison with Poison: Detecting Backdoor Poison Samples via Decoupling Benign Correlations

Figure 4 for Fight Poison with Poison: Detecting Backdoor Poison Samples via Decoupling Benign Correlations

Access Paper or Ask Questions

Circumventing Backdoor Defenses That Are Based on Latent Separability

May 26, 2022
Xiangyu Qi, Tinghao Xie, Saeed Mahloujifar, Prateek Mittal

Figure 1 for Circumventing Backdoor Defenses That Are Based on Latent Separability

Figure 2 for Circumventing Backdoor Defenses That Are Based on Latent Separability

Figure 3 for Circumventing Backdoor Defenses That Are Based on Latent Separability

Figure 4 for Circumventing Backdoor Defenses That Are Based on Latent Separability

Access Paper or Ask Questions

Formulating Robustness Against Unforeseen Attacks

Apr 28, 2022
Sihui Dai, Saeed Mahloujifar, Prateek Mittal

Figure 1 for Formulating Robustness Against Unforeseen Attacks

Figure 2 for Formulating Robustness Against Unforeseen Attacks

Figure 3 for Formulating Robustness Against Unforeseen Attacks

Figure 4 for Formulating Robustness Against Unforeseen Attacks

Access Paper or Ask Questions

ObjectSeeker: Certifiably Robust Object Detection against Patch Hiding Attacks via Patch-agnostic Masking

Feb 03, 2022
Chong Xiang, Alexander Valtchanov, Saeed Mahloujifar, Prateek Mittal

Figure 1 for ObjectSeeker: Certifiably Robust Object Detection against Patch Hiding Attacks via Patch-agnostic Masking

Figure 2 for ObjectSeeker: Certifiably Robust Object Detection against Patch Hiding Attacks via Patch-agnostic Masking

Figure 3 for ObjectSeeker: Certifiably Robust Object Detection against Patch Hiding Attacks via Patch-agnostic Masking

Figure 4 for ObjectSeeker: Certifiably Robust Object Detection against Patch Hiding Attacks via Patch-agnostic Masking

Access Paper or Ask Questions