publications

2026

1. arXiv

   Structured Thoughts For Improved Reasoning And Context Pruning

   Zain Sarwar, Supriyo Chakraborty, Berkcan Kapusuzoglu, and 4 more authors

   2026

   arXiv Bib HTML

   @misc{sarwar2026structuredthoughts,
     title = {Structured Thoughts For Improved Reasoning And Context Pruning},
     author = {Sarwar, Zain and Chakraborty, Supriyo and Kapusuzoglu, Berkcan and Lee, C. H. and Das, Anirban and Rawls, Stephen and others},
     year = {2026},
     archiveprefix = {arXiv},
   }

2. ICML

   Critique-Guided Distillation for Robust Reasoning via Refinement

   Berkcan Kapusuzoglu, Supriyo Chakraborty, Zain Sarwar, and 2 more authors

   2026

   arXiv Bib HTML

   @misc{kapusuzoglu2025critique,
     title = {Critique-Guided Distillation for Robust Reasoning via Refinement},
     author = {Kapusuzoglu, Berkcan and Chakraborty, Supriyo and Sarwar, Zain and Lee, C. H. and Sahu, Sambit},
     howpublished = {International Conference on Machine Learning (ICML)},
     year = {2026},
     archiveprefix = {arXiv},
   }

2025

1. NeurIPS

   

   Dense Backpropagation Improves Training for Sparse Mixture-of-Experts

   Ashwinee Panda, Vatsal Baherwani, Zain Sarwar, and 4 more authors

   In Advances in Neural Information Processing Systems (NeurIPS), 2025

   arXiv Bib PDF

   @inproceedings{panda2025dense,
     title = {Dense Backpropagation Improves Training for Sparse Mixture-of-Experts},
     author = {Panda, Ashwinee and Baherwani, Vatsal and Sarwar, Zain and Th{\'e}rien, Benjamin and Sahu, Sambit and Goldstein, Tom and Chakraborty, Supriyo},
     booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
     year = {2025},
   }

2. TMLR

   

   Continual Pre-training of MoEs: How robust is your router?

   Benjamin Thérien, Charles-Étienne Joseph, Zain Sarwar, and 7 more authors

   Transactions on Machine Learning Research (TMLR), 2025

   arXiv Bib PDF

   @article{therien2025continual,
     title = {Continual Pre-training of {MoEs}: How robust is your router?},
     author = {Th{\'e}rien, Benjamin and Joseph, Charles-{\'E}tienne and Sarwar, Zain and Panda, Ashwinee and Das, Anirban and Zhang, Shi-Xiong and Rawls, Stephen and Sahu, Sambit and Belilovsky, Eugene and Rish, Irina},
     journal = {Transactions on Machine Learning Research (TMLR)},
     year = {2025},
   }

2024

1. NeurIPS Workshop

   

   StructMoE: Structured Mixture of Experts Using Low Rank Experts

   Zain Sarwar, Ashwinee Panda, Benjamin Thérien, and 8 more authors

   In NeurIPS Efficient Natural Language and Speech Processing Workshop (ENLSP), 2024

   Bib HTML

   @inproceedings{sarwar2024structmoe,
     title = {{StructMoE}: Structured Mixture of Experts Using Low Rank Experts},
     author = {Sarwar, Zain and Panda, Ashwinee and Th{\'e}rien, Benjamin and Rawls, Stephen and Das, Anirban and Balasubramaniam, Kartik and Kapusuzoglu, Berkcan and Zhang, Shixiong and Sahu, Sambit and Naphade, Milind and Chakraborty, Supriyo},
     booktitle = {NeurIPS Efficient Natural Language and Speech Processing Workshop (ENLSP)},
     year = {2024},
   }

2. NeurIPS Workshop

   Dense Backpropagation Improves Routing for Sparsely-Gated Mixture-of-Experts

   Ashwinee Panda, Vatsal Baherwani, Zain Sarwar, and 4 more authors

   In Workshop on Machine Learning and Compression, NeurIPS, 2024

   Bib

   @inproceedings{panda2024densews,
     title = {Dense Backpropagation Improves Routing for Sparsely-Gated Mixture-of-Experts},
     author = {Panda, Ashwinee and Baherwani, Vatsal and Sarwar, Zain and Th{\'e}rien, Benjamin and Rawls, Stephen and Sahu, Sambit and others},
     booktitle = {Workshop on Machine Learning and Compression, NeurIPS},
     year = {2024},
   }

3. NeurIPS Workshop

   

   MYCROFT: Towards Effective and Efficient External Data Augmentation

   Zain Sarwar, Van Tran, Arjun Nitin Bhagoji, and 3 more authors

   In NeurIPS Workshop on Machine Learning for Systems (MLforSys), 2024

   arXiv Bib PDF

   @inproceedings{sarwar2024mycroft,
     title = {{MYCROFT}: Towards Effective and Efficient External Data Augmentation},
     author = {Sarwar, Zain and Tran, Van and Bhagoji, Arjun Nitin and Feamster, Nick and Zhao, Ben Y. and Chakraborty, Supriyo},
     booktitle = {NeurIPS Workshop on Machine Learning for Systems (MLforSys)},
     year = {2024},
   }

4. USENIX Security

   

   Can Virtual Reality Protect Users from Keystroke Inference Attacks?

   Zhuolin Yang, Zain Sarwar, Iris Hwang, and 3 more authors

   In 33rd USENIX Security Symposium (USENIX Security 24)Press: New Scientist , 2024

   arXiv Bib PDF

   @inproceedings{yang2024vr,
     title = {Can Virtual Reality Protect Users from Keystroke Inference Attacks?},
     author = {Yang, Zhuolin and Sarwar, Zain and Hwang, Iris and Bhaskar, Ronik and Zhao, Ben Y. and Zheng, Haitao},
     booktitle = {33rd USENIX Security Symposium (USENIX Security 24)},
     year = {2024},
   }

2023

1. USENIX Security

   

   Towards a General Video-based Keystroke Inference Attack

   Zhuolin Yang, Yuxin Chen, Zain Sarwar, and 3 more authors

   In 32nd USENIX Security Symposium (USENIX Security 23), 2023

   Bib PDF

   @inproceedings{yang2023keystroke,
     title = {Towards a General Video-based Keystroke Inference Attack},
     author = {Yang, Zhuolin and Chen, Yuxin and Sarwar, Zain and Schwartz, Hadleigh and Zhao, Ben Y. and Zheng, Haitao},
     booktitle = {32nd USENIX Security Symposium (USENIX Security 23)},
     year = {2023},
   }

2. IEEE S&P

   

   Deepfake Text Detection: Limitations and Opportunities

   Jiameng Pu^*, Zain Sarwar^*, Sifat Muhammad Abdullah, and 5 more authors

   In IEEE Symposium on Security and Privacy (S&P)Press: CACM, TechXplore , 2023

   arXiv Bib PDF

   @inproceedings{pu2023deepfake,
     title = {Deepfake Text Detection: Limitations and Opportunities},
     author = {Pu, Jiameng and Sarwar, Zain and Abdullah, Sifat Muhammad and Rehman, Abdullah and Kim, Yoonjin and Bhattacharya, Parantapa and Javed, Mobin and Viswanath, Bimal},
     booktitle = {IEEE Symposium on Security and Privacy (S\&P)},
     year = {2023},
   }

1. Preprint

   Low Rank Experts Enable Specialization In Dense Transformers

   Zain Sarwar, Ashwinee Panda, Benjamin Thérien, and 3 more authors

   Bib

   @misc{sarwar2025lowrank,
     title = {Low Rank Experts Enable Specialization In Dense Transformers},
     author = {Sarwar, Zain and Panda, Ashwinee and Th{\'e}rien, Benjamin and Naphade, Milind and Sahu, Sambit and Chakraborty, Supriyo},
   }