Abstract

TensorFlow is Google’s open source library for high performance numerical computation, based on a multi-stage API that constructs explicit computation graphs. In this talk, we will describe a new software library that automatically converts plain Python code into its TensorFlow equivalents, using source code transformation. For example, instead of tf.cond, and tf.while_loop, users would write imperative-style code with ifs and whiles, and expect it would be converted into its Graph equivalent. We support many Python idioms used in scientific programming, but explicitly opt to not support the language in its entirety. Our approach is complementary with the new TensorFlow Eager project and will allow using the imperative style of Eager mode, while retaining the benefits of graph mode. By using automatic code conversion, developers can write code that’s more concise, efficient and robust.

Bio

Alex Wiltschko is a senior research scientist at Google Brain. He obtained his PhD in neurobiology from Harvard Medical School, where he built new methods for understanding and parsing behavior and body language. He then co-founded Whetlab, a machine-learning-as-a-service company, which was acquired by Twitter in 2015. At Google Brain, Alex works on new tools for machine learning, as well as applying ML to impactful problems in biology and chemistry.

Alex Wiltschko is a senior research scientist at Google Brain. He obtained his PhD in neurobiology from Harvard Medical School, where he built new methods for understanding and parsing behavior and body language. He then co-founded Whetlab, an ML-as-a-service company, which was acquired by Twitter in 2015. At Google Brain, Alex works on new tools for machine learning, as well as applying ML to problems in biology and chemistry.