Research

Managing Data for Machine Learning

Data-centric machine learning calls for intelligently obtaining the best possible data for training a model. Data-centric practices can significantly reduce the financial, labor, and time costs of designing, training, and deploying AI systems in the wild. This research proposes operations-based approaches to data-centric modeling by optimizing what data to collect, synthesize, and label for building ML models.

Data-driven Optimization in Operations Research

Data-driven decision-making requires domain experts to manage data, tune models, and personalize decisions, all of which is laborious and can lead to delays in downstream operations. This research introduces new machine learning-based techniques for automated decision-making that can reduce this human labor and improve the quality of operational services.

Artificial Intelligence in Personalized Healthcare

Machine learning and data-driven optimization can automate healthcare services in order to improve effectiveness, quality of care, and access to treatment. This research applies predictive and prescriptive modeling in personalized medicine (e.g., cancer therapy and otolaryngology) and clinical services (e.g., milk bank donation management).

Real-time Multimedia Error Correction Coding and Information Theory

In real-time Internet streaming, data packets are encoded with error correction to protect against transmission failures as they flow through nodes in a network. This research introduces a series of error correction coding algorithms for real-time multimedia that ensures the data is resilient to the types of error patterns typically found in Internet streaming. I am no longer active in this area.