?? CSO Linkedin Spotlight Series

Artificial Intelligence and Machine Learning for Official Statistics Project (AIML4OS).

This project is a European Statistical System Network (ESSnet) collaborative project involving 16 member countries. The CSO (Central Statistics Office Ireland) is the project lead and co-ordinator for this exciting and innovative project which is being funded by the European Commission.

We recently sat down with Brendan O’Dowd and Eimear Crowley to chat about the CSO’s role in the AIML4OS project and what the function of the project is.

What is the AIML4OS project?

The AIML4OS project is a collaborative effort among European national statistical offices (NSOs) that will help those who produce official statistics to develop solutions that make use of AI and machine learning. It will do this by producing a comprehensive set of resources, guidelines, methodologies and use cases for NSOs to access. It kicked off in April of 2024 and will run for four years.

The work of the project is divided across 13 workpackages. We are in charge of the first of these, which is overall project management. There are another five workpackages with cross- cutting functions, and the remaining seven are use cases, each of which explores a specific tool or theme within AI/ML in the context of official statistics. For example, some of the use case workpackages examine the use of AI/ML for earth observation data, synthetic data and large language models.

What is the expected impact of the AIML4OS project?

Providing NSOs with a comprehensive set of resources to implement AI/ML solutions will help them to improve the quality, timeliness, and efficiency of their statistical production processes. It will allow them to make use of new data sources, and this might be accompanied by reduced burden for survey respondents or opportunities for new statistical products.

We also hope that the project will lead to the growth of networks and communities where people share their expertise, develop further collaborations, come up with cross-functional solutions and so on. One of the challenges for us as coordinators is to support the development of these communities and to try to stimulate future innovations. Already we are seeing the benefit of pooling our collective skills and experience, and since NSOs face so many similar shared challenges the rationale for working together is self-evident.

What is the role of the CSO in the AIML4OS project?

We are responsible for the overall coordination of the project, so we organise regular meetings among the 13 workpackages to ensure that all of the tasks are progressing as they should, and that each of the 39 deliverables that are due over the course of the four years are achieved on time. We oversee the finances of the project and liaise directly with Eurostat on any issues. We spend a lot of our time working on “horizontal” aspects, maintaining the many interconnections between workpackages, such as the assembly of materials in a knowledge repository, or ensuring that synergies between workpackages are exploited.

How is the CSO integrating AI/ML in it’s processes or outputs, presently?

Statistical Offices across Europe have been exploring the use of AI and ML for statistical production for many years now, in areas such as classification, data editing and imputation, typically to either improve efficiency or data quality, or to make use of novel sources of data. The pace of research has increased recently with the availability of technology. Our coordination role in the AIML4OS project provides us with an enviable central position within this space which allows us access to a lot of knowledge and expertise. One area where we are seeking to make use of AI is for text classification, where we try to use text descriptions to determine some categorisation, be it for products, occupations, or business sectors, for example. We have also had success in using GenAI for translating code from one statistical programming language to another, which has been met with a lot of interest from colleagues in NSIs around the world.

Want to learn more about the AIML4OS project?

??View the latest AIML4OS news: https://cros.ec.europa.eu/dashboard/aiml4os

??Subscribe to the AIML4OS newsletter - email: [email protected]

?? Follow the AIML4OS Linkedin page: https://www.dhirubhai.net/company/essnet-aiml4os/about/