Networks have become very complex over the past decade. The users and operators of large cloud platforms and campus networks have desired a much more programmable network infrastructure to meet the dynamic needs of different applications and reduce the friction they can cause to each other. This has culminated in the Software-‐defined Networking paradigm. But you cannot program what you do not understand: the volume, velocity and richness of network applications and traffic seem beyond the ability of direct human comprehension. What is needed is a sensing, inference and learning system which can observe the data emitted by a network during the course of its operation, reconstruct the network's evolution, infer key performance metrics, continually learn the best responses to rapidly-‐changing load and operating conditions, and help the network adapt to them in real-‐time. The workshop brings together academic and industry groups interested in the broad themes of this topic. It highlights ongoing research at Stanford and describes initial prototype systems and results from pilot deployments.