The Hackathon for Peace Justice & Security in The Hague has been won by Open DSE. They developed a solution, named Faked Sniff, to detect fake video’s (e.g. face swaps and deepfakes). The new location of The Hague Tech in the former Ministry of Social Affairs (SOZA) building was the setting for the international Hackathon for Peace, Justice & Security this weekend. The start of the hackathon coincided with the kick-off of the IT & Security Hub in The Hague, the innovative heart of the city where more than 125 organistions work on tech and cybersecurity solutions for a better world.
Over 200 tech-talents participated in the second edition of the Hackathon for Peace, Justice & Security - organized by the Municipality of The Hague in collaboration with The Hague Tech. The jury chose Open DSE as the overall winner & Challenge winner because, despite the difficult challenge, they developed a simple-to-use app to detect deepfake video’s. The team won a cash prize of € 5,000. In addition, 4 prizes of € 2,000 were available for the best solutions to challenges from Translators without Borders, 510 Red Cross as well as the European Commission & Space4Good.
Doing Business and Doing Good
Saskia Bruines, Deputy Mayor of The Hague: “This event is a good example of both doing business and doing good in The Hague. More and more organsiations work together in the IT & Security Hub of The Hague on solutions to global problems. In the coming years, we want to give this area an extra boost, attract more tech-talent to the area, and we hope to organise this great tech event more often.”
Winners Hackathon for Peace, Justice & Security
UNICRI Centre for Artificial Intelligence and Robotics
Overall/Challenge winner: Open DSE
With the team’s solution, named Faked Sniff, they contribute to tackling the problem of fake videos. With the rise of digital technology, fake videos have become a powerful propaganda tool. It seems the human brain is not very good at determining whether a video is fake. Open DSE cleverly integrated the power of AI to help find determining fake videos. By the newest, state of the art AI based on neural networks that are trained with large datasets, fake videos can be detected, mainly by focusing on facial movements.
Translators without borders (TWB)
Winning team: The lazy panda’s
Translators Without Borders has a huge database with relevant information. The key question is: which documents are in urgent need of translation in which are less urgent? The problem is the lack of relevant category information in the database. The team developed an algorithm that uses social media to help categorizing and assessing the impact of the document. Adding valuable outside information from social media platforms in the relevant country helps determining the relevance and impact of the documents. However the algorithm is a valuable contributor, the team’s solution is interactive and offers the option for human intervention in re-assessing the impact of documents.
510 Red Cross
Winning team: Ellipsis Earth Intelligence
This team worked on getting and validating info for aid after disasters. Take for example, the earthquake in September 2018 in Indonesia. Quickly after the disaster, many question arose: what is the damage? How many victims are there and where are they located?
Only 6 hours after the disaster, the Red Cross data team has gathered loads of information from satellites, spectral imaging, weather forecasts and other sources. This information is shared on an online platform. The team developed an app that accesses the data on this platform. The app shows locations that are in need of a visit, based on dame, predicted number of victims, etc.
Europese Commissie & Space4Good
Winning team: Bootleg against poverty
The team developed Enershare, an energy sharing app. In their philosophy, sharing energy shouldn’t cost energy. Solar energy often comes from small communities. Sometimes it’s cloudy, sometimes it is sun, so what to do with the surplus energy? Let’s share it. Let’s create a marketplace with prosumers and consumers. What does it mean at the human level? Prosumers (who generate the solar energy) don’t want too make too much effort, keep their autonomy and see their impact. The app developed by the team manages this surplus, giving prosumers the choice to whom to sell it and at what price.
National Rapporteur reports on Trafficking in Human Beings and Sexual Violence
Winning team: CGOA
The team tackled the challenge of the huge library of the National Rapporteur. How do stakeholders find the relevant information on their website? Now there are too many questions and disappointed website visitors. The team developed a new search engine: with an improved algorithm and a better user interface. As a visitor you can now key in search terms and the site shows you the publication dates and relevancy of the documents. If you click on the documents you can narrow down the search with additional search terms and share query’s with other people online. Any data library with large amounts of pdf’s (e.g. from universities, government bodies) can make use of this technology.