Raghav Somani

Papers

Total Citations

H-Index

About

Raghav Somani is a researcher whose work lies at the intersection of meta-learning and robust statistics, with a particular focus on enabling effective learning from limited data. His primary research areas include meta-learning, mixed linear regression, and robust optimization in low-data regimes. Somani’s major contributions address a critical challenge in modern supervised learning: how to train models effectively when each task provides only a small number of labeled examples—a common scenario in fields like medical image processing and robotic interaction. In his most cited work, "Robust Meta-learning for Mixed Linear Regression with Small Batches" (2020, 6 citations), he developed methods that exploit task similarities to overcome data scarcity, even when tasks are heterogeneous. His follow-up paper, "Meta-learning for Mixed Linear Regression" (2020, 2 citations), further formalized this approach. While his citation counts are modest, Somani’s work is notable for tackling a fundamental bottleneck in practical machine learning: the tension between task diversity and limited per-task data. His research offers promising pathways for deploying AI in data-constrained, real-world environments.