Amardeep Sharma

Title: Amardeep Sharma and Machine-Learning Evaluation for Feature Bias.

Introduction

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

Latest Patents

The provided information lists one patent for Sharma: “Machine learning evaluation for detecting feature bias.” The invention describes an evaluation framework that uses evaluation data structures to assess feature bias in machine-learning models. It receives an evaluation dataset containing different classes of an evaluation feature, generates at least two performance metrics, and calculates a unitless dissimilarity metric using those metrics and their average. This metric reflects variability associated with the evaluation feature. The patent count provided is 1 patent.

Career Highlights

Sharma is associated with Optum, Inc. His listed patent highlights work on methods for evaluating machine-learning performance and detecting potential bias in model features.

Collaborations

Rahul Pathak is identified as Sharma’s coworker. The provided information does not specify whether Pathak collaborated on the patent.

Conclusion

Amardeep Sharma’s listed invention presents a structured approach to measuring feature bias in machine-learning models, using performance metrics and a unitless dissimilarity measure.