Julia's well-known combination of ease-of-use, performance and powerful features make it uniquely suited to the toughest machine learning problems. We'll illustrate how Julia can accelerate your current workflow, show you the groups running intelligent Julia code in production, and discuss our plans for the future.
Existing machine learning frameworks are complex "black boxes" which are designed to work at a certain level of abstraction. If you need higher-level (e.g. complex models) or lower-level (e.g. custom gradients or GPU kernels) control than the framework provides, you get stuck.
Julia's mix of performance and ease of use opens a radical alternative; a single language that can work at all levels, from GPUs to data processing pipelines to clusters. Different approaches can be freely mixed and matched, with transparency and control at all levels of the stack. Come see how these ideas can accelerate your current workflow, and get a glimpse of the future of ML in Julia.