Abstract: With the increasing use of AI in high-stakes domains such as medicine, law, and defense, organizations spend a lot of time and money to make ML models trustworthy. Many books on the subject offer deep dives into theories and concepts. This tutorial provides a practical starting point to help ML development teams produce models that are secure, more robust, less biased, and more explainable.
Drawing from our collective expertise and experience as trustworthy ML practitioners, we translate best practices in the academic literature for curating datasets and building models into a blueprint for building industry-grade trusted ML systems using a series of deep-dive, hands-on examples. All code used in this tutorial will be available on GitHub.
Bio: Subho Majumdar is a technical leader in applied trustworthy machine learning who believes in a community-centric approach to data-driven decision making. He has pioneered the use of trustworthy ML methods in industry settings, co-wrote a book, and founded multiple nonprofit efforts in this area. In past work, he has helped drive policy changes in government and nonprofit organizations through successful collaborations in the data for good space. Subho has 10 years of R&D experience in ML, data science, and statistics, with 30+ publications and 15+ filed patents (2 granted). Currently, Subho is a senior scientist in the Security ML group of Splunk. He holds a PhD and masters in statistics from University of Minnesota.