Mansi Sood

Research

I am currently a Schmidt Science Fellow at MIT Laboratory for Information and Decision Systems.

My research focuses on foundations for learning and inference in complex networked systems. The questions I study arise from operational demands in emerging computational paradigms, such as efficiently learning from limited, distributed, or dependent data, and from societal-scale challenges, such as evaluating policy interventions during epidemics. I develop scalable and interpretable inference algorithms that leverage graph structure to build mechanistic insight into system behavior and inform reliable decision-making.

Salient Themes/Publications

My current research has contributed methods for reliable inference in heterogeneous, interconnected, and partially observed systems.

Learning and Intervening under Latent Heterogeneity

M. Sood, D. Shah, Causal Inference with Unobserved Confounding: a Mixture Learning Perspective, IEEE Signal Processing Magazine: Special Issue on Methodological Advances in Causal Inference (under final review), 2026.
M. Sood, D. Shah, Learning Mixtures of Exponential Family, IEEE International Symposium on Information Theory, 2026.

Modeling and Mitigating Spreading Processes

M. Sood, H. Gu, R. Eletreby, S. Kumar, C. W. Wu, O. Yagan, On the Interplay of Clustering and Evolution in the Emergence of Epidemic Outbreaks, IEEE/ACM Transactions on Networking (accepted), 2026.
M. Sood, A. Sridhar, R. Eletreby, C. W. Wu, S. A. Levin, H.V. Poor, O. Yagan, Spreading Processes with Mutations over Multi-layer Networks, Proceedings of the National Academy of Sciences, 2023.

Random Graph Foundations for Distributed Inference

M. Sood, E. C. Elumar, O. Yagan, On Balancing Sparsity with Reliable Connectivity in Distributed Network Design with Random K-out Graphs, 2025, preliminary results in IEEE International Conference on Communications (ICC) 2021. [Best Paper Award]
M. Sood, O. Yagan, Existence and Size of the Giant Component in Inhomogeneous Random K-out Graphs, IEEE Transactions on Information Theory 2023.
M. Sood, O. Yagan, On the Minimum Node Degree and k-connectivity in Inhomogeneous Random K-out Graphs, IEEE Transactions on Information Theory 2021.

PhD Dissertation

M. Sood, Structural Heterogeneity and Performance in Stochastic Networks: From Distributed Inference to Epidemics and Beyond, PhD Thesis, Carnegie Mellon University, Aug, 2024. [A. G. Jordan Award for Outstanding Thesis] [ITA Graduation Day Award]

Google Sites

Report abuse