Datasets:
| { | |
| "field_id": "F021", | |
| "field": "automated machine learning", | |
| "directions": [ | |
| { | |
| "direction_id": "G0", | |
| "name": "Automated design and selection of predictive solutions", | |
| "scope": "Automatically find effective learning-algorithm configurations or predictive solutions while accounting for search and evaluation costs. This fully encompasses hyperparameter optimization, recommending candidates using evidence from prior tasks, constructing multistage pipelines, feature engineering, and discovering model structures and design components." | |
| }, | |
| { | |
| "direction_id": "G1", | |
| "name": "Diagnosis of AutoML search behavior", | |
| "scope": "Systematically study when automated search is beneficial, how its performance varies across problems, and how configuration spaces and optimization methods behave." | |
| }, | |
| { | |
| "direction_id": "G2", | |
| "name": "Practical and human-centered AutoML use", | |
| "scope": "Study how to bring AutoML into usable predictive-modeling workflows and support human participation, tool adoption, and continued development and operation in real-world settings." | |
| } | |
| ] | |
| } | |