Machine Learning Project Proposal BSTN ML team November

Machine Learning Project Proposal BSTN ML team, November 2018 .

Context • Machine learning is here to stay – Or broader: Algorithmic Approaches/Artificial Intelligence • ‘Everybody’ is struggling with the right application – EU communication on AI – Dutch Digital Strategy – Autonomous driving – Commercial recommender systems • Challenges are worldwide the same • Requires combination of different skills – …and don’t forget the role of data used! • Statistical offices are all on the same page • Excellent time to launch HLG-MOS project on ML ü Increasing public and private investments ü Preparing for socioeconomic changes ü Appropriate ethical and legal framework

ML position paper (4/4) • Recommendations Ø Develop quality framework v New, or integrated in existing one? v Possibly als guidelines for ML use Ø Create inventory of ML projects v To exchange both methods and applications v Not one-off but living document v Including previous work (Chu&Poirier, Destatis) Ø Set up interdisciplinary teams for ML work v Modelers v Programmers v subject matter specialists

ML project proposal • Proposed work packages 1. ML pilots 2. Inventory of existing ML projects 3. Design of a quality framework 4. Development of a ‘ML for dummies’ handbook 5. Communication and dissemination 6. Overall project management • All WPs can run in parallel, some dependencies • Most work virtual, 2 -3 physical sprints • Interdisciplinary team • Agile approach (identify POs? ) • Sandbox environment for pilots

Relation with HLG-MOS mission • The HLG-MOS (…) mission is to work collaboratively to identify trends, threats, and opportunities in modernising statistical organisations. It provides a common platform for experts to develop solutions in a flexible and agile way (…) • Four Vision elements Ø Actively engage Ø Be a trusted data authority Ø Adopt a service oriented approach Ø Have an agile adaptive culture • Five Values Ø Developing innovative solutions Ø Demonstrating leadership and collaboration Ø Discussing challenges and opportunities openly Ø Ensuring that priorities are community driven Ø Supporting a flexible, results oriented and agile approach

Relation with HLG-MOS mission üThe HLG-MOS (…) mission is to work collaboratively to identify trends, threats, and opportunities in modernising statistical organisations. It provides a common platform for experts to develop solutions in a flexible and agile way (…) üFour Vision elements ü Actively engage ü Be a trusted data authority ü Adopt a service oriented approach ü Have an agile adaptive culture üFive Values ü Developing innovative solutions ü Demonstrating leadership and collaboration ü Discussing challenges and opportunities openly ü Ensuring that priorities are community driven ü Supporting a flexible, results oriented and agile approach