Urvi Sharma

Title: Urvi Sharma: Advancing Machine-Learning Evaluation

Introduction

Urvi Sharma is an inventor associated with Uttar Pradesh, India. Her patent work addresses the evaluation of machine-learning models, with a focus on identifying feature bias.

Latest Patents

Sharma is associated with one patent, titled “Machine Learning Evaluation for Detecting Feature Bias.” The disclosed approach uses an evaluation framework and evaluation data structures to assess feature bias in machine-learning models. It receives an evaluation dataset containing different feature classes, generates at least two performance metrics, and uses those metrics and their average to calculate a unitless dissimilarity metric. This metric reflects variability associated with the evaluation feature. Sharma’s reported portfolio includes 1 patent.

Career Highlights

Sharma’s listed patent work is connected to Optum, Inc. The invention focuses on making model evaluation more comprehensive by providing a way to measure differences in performance across feature classes.

Collaborations

Amardeep Sharma is identified as a coworker. The available information does not specify her role in the patent or the invention.

Conclusion

Urvi Sharma’s patent describes a structured method for evaluating feature bias in machine-learning models, contributing to efforts to assess model performance across different feature classes.