Open source libraries for trusted ML and open standards for model deployment


Machine learning applications are now appearing in more and more places, sometimes affecting decision making in such critical areas as finance, medicine, or criminal justice. This has brought the issues of trust, fairness, explainability to the forefront of ML research. We will discuss open source libraries Adversarial Robustness 360 toolbox, AI Fairness 360, AI Explainability 360 created by IBM Research and now accepted to Linux Foundation Trusted AI committee. In addition, an important question is how to easily deploy, store, exchange machine learning models. Open standards help to break barriers between different commercial products and open source packages in terms of model exchange and deployment. We will talk about PMML, PFA, and ONNX. We will have hands-on exercises running Jupyter notebooks with the above libraries inside cloud based Watson Studio, as well as exporting PMML (and possibly ONNX) from some ML models and deploying it into Watson Machine Learning.

Our speaker:

Svetlana Levitan, PhD, is a Developer Advocate with Center for Open Source Data and AI Technologies (CODAIT) at IBM. She has been a software engineer, architect, and technical lead for SPSS Analytic components for many years. She represents IBM at the Data Mining Group and is the release manager for PMML and PFA, open standards for predictive model deployment. She is also working with other companies on ONNX, an open model exchange format for deep learning models. Svetlana is a co-organizer of several Chicagoland Meetup groups. She has authored several blogs and presented at many conferences and other events. Svetlana loves to learn new technologies and to share her expertise, to encourage girls and women in STEM.