Momin Ahmad Khan

Knowles Engineering Building
151 Holdsworth Way
Amherst, MA 01003
I am a fourth-year PhD student at UMass Amherst, working with Professor Fatima Anwar. My areas of interest are the security and privacy of distributed machine learning systems. I have worked on designing new attacks and defenses for Federated Learning (FL) and uncovering experimental pitfalls in the robustness evaluations of existing attacks and defenses.
Currently, I am a research intern at Nokia Bell Labs in the Autonomous Systems Research Group, where I am working on mechanistic interpretability for embodied AI. In the summer of 2024, I was a research intern in the same group at Bell Labs. I worked on designing model selection and code generation pipelines using various LLM agents and tools. Additionally, I developed a smart meeting manager utilizing LLMs, computer vision models, and hardware components to automate the orchestration and conduct of meetings without human intervention.
Before joining UMass, I completed my undergraduate studies in Electrical Engineering at the School of Electrical Engineering and Computer Science at the National University of Sciences and Technology in 2021. I graduated with a gold medal for best thesis project and a silver medal for the second-highest GPA in my batch.
Besides work, I am learning to play the guitar. I enjoy cooking and continue to try out new recipes. But most of all, I love playing Dota 2, and I am in the top 10% of players globally!
news
Jul 8, 2025 | Our new preprint Decoding FL Defenses: Systemization, Pitfalls, and Remedies is out! |
---|---|
Jun 10, 2025 | Our new preprint SABRE-FL: Selective and Accurate Backdoor Rejection for Federated Prompt Learning is out! |
Jun 2, 2025 | I returned to Nokia Bell Labs as a Research Intern |
Sep 25, 2024 | Our work, Hybrid Knowledge Distillation for Robust and Accurate FL (HYDRA-FL) is accepted to Neurips 2024! |
Aug 23, 2024 | I completed my internship at Nokia Bell Labs as part of the Autonomous Systems Research Group! |
selected publications
- NeurIPSHYDRA-FL: Hybrid Knowledge Distillation for Robust and Accurate Federated LearningIn 38th Conference on Neural Information Processing Systems (NeurIPS 2024), 2024
- Under-review
- SenSys
- IROSA Neurosymbolic Approach to Adaptive Feature Extraction in SLAMIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024
- IEEE DLSPOn the Pitfalls of Security Evaluation of Robust Federated LearningIn 2023 IEEE Security and Privacy Workshops (SPW), 2023
- AIChallengeIOTSecurity Analysis of SplitFed LearningIn Proceedings of 4th International Workshop on Challenges in Artificial Intelligence and Machine Learning for Internet of Things (AIChallengeIoT 2022), in conjunction with ACM SenSys 2022, 2022