Let the Models Speak: Model Interpreters
Curiosity kills. We have advanced to building machine learning algorithms for real-time predictions that can help us predict almost accurately every time. It’s like we have built a magic box that will answer all our questions. The more curios question is why did the magic box predict what it did. During the session I would like to share why you should/shouldn’t trust your model but you should explain the reasoning of your predictions that help Business users to understand your Magic box. I would like to share about how at AT&T we are using Model Interpreters to help business users identify the elements of the predictions that can give intuitive reasoning behind each prediction.
Sandeep David Golkonda is a Data Scientist at AT&T Chief Data Office he actively researches and applies Machine Learning and AI solutions for real-time analytics problems. His, current work is focused on building AI framework to optimize network events. He is an active member of School of AI and Teaching Assistant for machine learning and deep learning. He holds M.S in computer science and M.S in business analytics.