Rajitha Meka

About

Rajitha Meka is a researcher advancing the frontiers of efficient global optimization for expensive computer experiments. Her primary research areas lie at the intersection of statistical machine learning, Bayesian optimization, and experimental design, with a particular focus on developing algorithms that balance exploration and exploitation under tight computational budgets. Meka’s most notable contribution is the introduction of the Multi-Armed Bandit Regularized Expected Improvement (BREI) method, a novel framework that integrates multi-armed bandit strategies with traditional expected improvement criteria. This approach significantly enhances the efficiency of optimizing black-box functions where each evaluation is costly, such as in complex physical simulations. By regularizing the search process, BREI reduces the risk of premature convergence and improves robustness in low-noise environments. Although her work is early-stage, with 2 citations to date, the BREI method represents a promising step forward in making global optimization more practical for real-world engineering and scientific applications. Meka’s research is particularly valuable for students and practitioners seeking to push the boundaries of sample-efficient optimization in high-stakes, resource-constrained settings.

Research Focus

Key Achievements

Top Papers

1. 1
   Multi-Armed Bandit Regularized Expected Improvement for Efficient Global Optimization of Expensive Computer Experiments With Low Noise

   2 citations · 2021

Key Collaborators

Contact & Links