Raman Arora

Raman Arora is an Associate Professor in the Department of Computer Science at Johns Hopkins University, with affiliations to the Mathematical Institute for Data Science (MINDS), the Center for Language and Speech Processing (CLSP), and the Institute for Data-Intensive Engineering and Science (IDIES). His research spans theoretical and practical aspects of machine learning, focusing on robustness, privacy, representation learning, and optimization.

Research Interests:

• Machine Learning Theory
• Representation Learning (e.g., Deep CCA, Multi-view Learning)
• Privacy-Preserving Machine Learning (Differential Privacy)
• Robustness in Deep Learning
• Online and Reinforcement Learning
• Stochastic Optimization Algorithms

His recent publications, primarily in top-tier venues like NeurIPS, ICML, and ICLR, demonstrate a strong focus on the theoretical foundations of adversarial robustness, multi-task learning, offline reinforcement learning, and differentially private optimization. His work often bridges theory and practice, with applications in speech, language, and data-intensive systems.

Scientific Awards and Honors:

• NSF CAREER Award (2020)
• ICML Test-of-Time Award Finalist (2023) for Deep CCA
• Member, Institute for Advanced Study (2019–2020)
• Visiting Scientist, Simons Institute (2019, 2020, 2022)

Advising and Grants: Raman Arora has advised numerous PhD and master’s students, many of whom are now researchers at leading tech companies like Google, Meta, and Microsoft. His research is supported by significant grants from the NSF (including CAREER, BIGDATA, TRIPODS, and CRCNS awards), DARPA, and other agencies, focusing on foundational aspects of machine learning such as inductive biases, privacy, robustness, and computational neuroscience.

Laboratory and Research Group: He leads a dynamic research group at Johns Hopkins, comprising current PhD students and postdoctoral researchers working on the intersection of theory and applications in machine learning. The group is actively involved in projects related to adversarial robustness, meta-learning, offline reinforcement learning, and private optimization.