1: \begin{abstract}

2: 

3: The convergence of HPC and data intensive methodologies provide a promising

4: approach to major performance improvements. This paper provides a general

5: description of the interaction between traditional HPC and ML approaches and

6: motivates the "Learning Everywhere" paradigm for HPC. We introduce the concept

7: of "effective performance" that one can achieve by combining learning

8: methodologies with simulation based approaches, and distinguish between

9: traditional performance as measured by benchmark scores. To support the

10: promise of integrating HPC and learning methods, this paper examines specific

11: examples and opportunities across a series of domains. It concludes with a

12: series of open computer science and cyberinfrastructure questions and

13: challenges that the Learning Everywhere paradigm presents.

14: 

15: \end{abstract}

16: