Read about what happened with vera.ai’s first year, what the project has accomplished and where is it heading Take a look with us at 365…
It’s been a year since the VIGILANT consortium gathered for the project’s kick-off meeting when 17 European partners spanning academia, industry, government research centers, and Police Authorities (PAs) gathered together in Dublin, Ireland. The aim of the project is to help PAs detect and analyze disinformation from multiple online sources, modalities, and languages.
The project has come a long way since then, so let’s take a look at some highlights as well as Ontotext’s involvement.
The first few months of VIGILANT were packed with numerous workshops to establish the PAs’ needs and use cases as well as the ethical framework of the project.
The 4 PAs that took part presented a lot of similarities, but also notable differences in the types and sources of disinformation they faced. Consequently, their workflows and approaches also overlapped and diverged. So, the proposed solution had to be flexible and adaptable to the various relevant scenarios. As the PA’s infrastructure varied as well, the solution also had to be modular to allow scaling up and down of different components based on specific needs and limitations.
From the start, the consortium of partners has demonstrated a strong commitment to compliance with EU’s current and upcoming regulations. Notably, while all consortium members must comply with GDPR (General Data Protection Regulation), some articles and limitations apply to academic and industry partners, while others apply to PAs. Therefore, the solution had to address all relevant regulations. These considerations and commitment to ethics were solidified in two deliverable reports produced six months after the start of the project. They described the consortium’s approach to data protection, privacy, and Artificial Intelligence.
Shortly after these events came the 2023 GLOBSEC forum, held as usual in Bratislava, Slovakia. Ontotext was among the partners invited to present the project to an audience of PAs and other Law Enforcement Agencies (LEAs). This proved to be a great opportunity to gain more insights on PAs’ needs. In addition, LEAs were encouraged to join the project’s Community of Early Adopters to receive regular updates on the project’s progress and participate in further workshops.
However, VIGILANT is not the only project helping LEAs address the growing problem of disinformation. That is why its development should not happen in a vacuum. Instead, the project should learn from and share with other initiatives addressing similar issues. To that effect, VIGILANT participated in multiple EU Research events organized by CERIS and REA, as well as by EU-HYBNET and other projects in the field.
The consortium did not just learn from such events. It was also busy sharing findings. Among the project’s 2023 scientific publications is a paper on event detection by Ontotext that was presented at the CASE (Challenges and Applications of Automated Extraction of Socio-political Events from Text) workshop at the 2023 RANLP (Recent Advances in Natural Language Processing) conference.
In September 2023, the consortium gathered in Bratislava for its 3rd General Assembly to define the scope of the project’s Minimum Viable Product (MVP) and outline the activities for the next year. The technical partners are now working intensively on completing the MVP by the middle of the second year.
Ontotext in particular is working on ingestion, storage and enrichment of data. While most of the data sources will be online, we are also enabling users to add and interlink data from other sources, such as archives.
The data to be stored comes from multiple sources, modalities and languages. Ontotext is responsible for designing a harmonized data model that allows for integration of data not just from the current sources of interest, but also future ones. Similarly, the data model has to support multiple modalities, including text, audio, image, video, etc.
Ontotext will also contribute to the enrichment of the collected data via multiple services, including language tagging, entity linking and event detection. These and other enrichments provided by the project partners will enable the better exploration and understanding of the collected data. To this end, Ontotext is tasked with the configuration of informative facets based on the enrichments that allow the user to effectively filter the collected information.
Half a year from now, the PAs will finally get to experience the first version of the system being developed. Their feedback will provide information to the technical partners on how to best approach the final version of the system. It is to be provided in the third year of the project along with extensive training and training materials.
The consortium will also continue its external engagement efforts in the second and third years. Multiple workshops with the broader general public, other LEAs, other adjacent projects and other project stakeholders have already been planned and scheduled. VIGILANT will keep engaging a wide range of stakeholders and address their needs, questions, and concerns.
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No:101073921. Views and opinions expressed are however those of the author only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.