Our way forward
People often come to us and ask what SAP does in the area of machine learning, and how we keep on top of this fast-paced field. There is a clear answer to these questions — we build new and empower existing products with machine learning and we do continuous research.
Our teams spend considerable time working on tough machine learning problems and techniques digging deeper into diverse areas, such as computer vision or natural language processing, where machine learning in general and deep learning in particular are changing the rules of the game. And we believe that these activities should be made visible — our goal is to bring them back to the community by providing insights into what type of machine learning problems we encounter in the enterprise world and how we try to solve them.
As a research team, we work together with customers, SAP product teams, and the SAP Deep Learning Center of Excellence. From these groups, we get the three most important ingredients that make our job so exciting: challenging problems, rich data, and a hard to beat baseline. Our goal is to carry out top of the line research to improve the existing baseline and to spot parallels across different problems. We also work on novel approaches and machine learning fields we believe will be important in the future.
Who are we? We are SAP’s machine learning researchers spread over Berlin, Singapore, Walldorf, Palo Alto and Tel Aviv. Our team ranges from data scientists to engineers to PhD students. We work together with our academic partners, such as MIT, Stanford or University of Amsterdam in order to collaborate with the brightest minds.
You can consider this blog as our machine learning notebook, where we keep most interesting findings from our ongoing research projects, provide highlights from different conferences and post small and big victories, as well as failures. We would also like to share our thoughts on different machine learning topics, such as domain adaptation, few shot learning, or efficient deep learning, which we consider relevant for building enterprise software.
We know machine learning has the potential to transform all business areas. However, through our daily work we are also well aware of its current limits and that there is still a long way to go. We are looking forward to sharing our experience with you through this blog — in the hope that we can learn from each other. Feel free to stop by and let us know what you think!
