As the world moves to using machine learning more and more, a question arises of how to prevent cyber criminals from attacking systems using ML frameworks. In this talk, we will cover why machine learning needs security practices and how we develop TensorFlow to be secure. We will discuss topics such as fuzzing and code transformations for secure software development with a focus on how these are applied within TensorFlow ecosystem.
As the world moves to using machine learning more and more, a question arises of how to prevent cyber criminals from attacking systems using ML frameworks. In this talk, we will cover why machine learning needs security practices and how we develop TensorFlow to be secure. We will discuss topics such as fuzzing and code transformations for secure software development with a focus on how these are applied within TensorFlow ecosystem.