Acknowledgement
Machine Learning (ML) holds significant potential for enhancing efficiency by complementing or replacing traditional methods, as well as improving quality in ways that are challenging for traditional methods to achieve. However, numerous barriers hindering statistical organisations from implementing ML methods for editing are non-methodological. The objective of this document is to identify these barriers and offer guidance on avoidance or overcoming them. With this document, we aim to encompass discussions and advice on addressing the identified issues, along with a compiled set of use cases gathered during the task team work.
We thank the following team members for their generous dedication of time and for contributing their valuable input:
Claire Clarke (chair) and Jenny Pocknee –Australian Bureau of Statistics
Wesley Yung, Jean Le Moullec and Stan Hatko –Statistics Canada
Riita Piela –Statistics Finland
Steffen Moritz – German Federal Statistical Office
João Poças –Statistics Portugal
Sandra Barragán, David Salgado and Elena Rosa-Pérez –Statistics Spain
Jens Malmros –Statistics Sweden
Daniel Kilchmann – Swiss Federal Statistical Office
Olivier Sirello and Bilyana Bogdanova – BIS
Amilina Kipkeeva – UNECE