Machine learning systems are stuck in a rut. Paul Barham and Michael Isard, two of the original authors of TensorFlow, come to this conclusion in their recent HotOS paper. They argue that while TensorFlow and similar frameworks have enabled great advances in machine learning, their current design and implementations focus on a fixed set of monolithic and inflexible kernels. They continueto say that “this reliance on high performance but inflexible kernels reinforces the dominant style of programming model” and argue that “these programming abstractions lack expressiveness, maintainability, and modularity; all of which hinders research progress”.
|Number of pages||6|
|Publication status||Published - 31 May 2020|
|Event||2nd Workshop on Accelerated Machine Learning @ ISCA 2020 - Virtual workshop|
Duration: 31 May 2020 → 31 May 2020
|Workshop||2nd Workshop on Accelerated Machine Learning @ ISCA 2020|
|Abbreviated title||AccML 2020|
|Period||31/05/20 → 31/05/20|