D3M 2017 - Data-driven Discovery of Models (D3M)

Contributions describing work in progress as well as position papers are invited. All contributions must focus on data-driven approaches to the design and deployment of ML/DM models. Of particular interest are methods and proposals that address the following issues:
Ensemble learning
Learning with privileged information
Learning with missing data
Learning from heterogeneous data
Metalearning
Hyperparameter optimization
Non-parametric, model-free, and zero-knowledge learning
Non-parametric causal inference
Spectral graph embedding and inference
Automatic feature generation
Weak supervision
Workflows with multiple learning methods
Planning and optimizing learning workflows