Traditionally, software development infrastructure leverages heuristics or arbitrary decisions to optimize performance. However, research is increasingly identifying areas where machine learning can help achieve better-than-human performance in the software domain. In the development of software code, potential machine learning applications include neural program synthesis (where code is generated automatically) or neural decompilation. During the build process, machine learning can be used to automatically optimize code or to perform fuzz testing, an automated testing approach to identify program exceptions like crashes or memory leaks. And finally, once code has been deployed, machine learning can be used for automated trace debugging and/or configuration management.
Gideon Mann is the head of Data Science at Bloomberg, where his team works on the company-wide data science platform, natural language question answering, and deep learning text processing, among other products. He also founded and leads the Data for Good Exchange (bloomberg.com/d4gx), an annual conference on data science applications for social good. Mann graduated Brown University in 1999 and subsequently received a Ph.D. from The Johns Hopkins University in 2006. After a short post-doc at UMass-Amherst, he moved to Google Research in NYC in 2007. In addition to academic research, his team at Google built core internal machine learning libraries and publicly released the Google Prediction API and Colab, a collaborative iPython application. He joined Bloomberg's leadership team in the CTO department in 2014.