NEW YORK, 29 May 2019 (Office of Information and Communications Technology) - The Secretariat of the United Nations High-Level Committee on Management (HLCM) today announced that a team from the University of Bologna, Italy, won a global challenge for the Extraction and Elicitation of United Nations General Assembly Resolutions. The University of Bologna team was formed by Monica Palmirani, Fabio Vitali, Silvio Peroni, Aldo Gangemi, Andrea Nuzzolese, Octavian Bujor, Biagio Distefano, Francesco Draicchio, Davide Liga and Francesco Sovrano.
This challenge was organized by the United Nations High-Level Committee on Management (HLCM) of the United Nations Chief Executives Board for Coordination (CEB) and was hosted on “Unite Ideas”. Unite Ideas is an open innovation and crowdsourcing platform provided by the United Nations Office of Information and Communications Technology. It is intended to empower computer scientists and innovators worldwide from academia, civil society and the private sector to tackle complex issues faced by the Organization and its members.
The challenge aimed to engage computer scientists to produce a programme capable of analysing official United Nations Documents and extracting and tagging entities and structures.
The challenge is part of the ongoing efforts by HLCM and its lead partners, such as the UN Department of General Assembly and Conference Management (DGACM) and the Food and Agriculture Organization, to produce documents in the machine-readable standard Akoma Ntoso forthe United Nations (AKN4UN), the adopted UN standard for parliamentary, legislative and judiciary documents. These enhanced documents would improve accessibility and retrieval of information for Member States, UN Staff, researchers and policymakers worldwide, supporting informed and efficient decision-making.
The winning team made available an open source tool called SANKOFA (Semantic Annotation of Knowledge Featuring Akoma Ntoso), capable of formatting United Nations resolutions and extracting knowledge from the text by employing natural language processing techniques. SANKOFA can detect the structure of a document, its references, as well as the names of the persons, the places and the organizations mentioned in it, among other attributes.
The Judging Panel also commended the work of a second team that participated in the challenge, that from Sheffield University, formed by Adam Funk, Kalina Bontcheva, Mark A. Greenwood, and Ian Roberts. Their solution, Open-source text processing with GATE, based on the open source software toolkit GATE, was considered very promising and worthy of further development.