Ministry of Economy: Over 50 IT specialists in Kyiv trained AI to accelerate demining

Thirteen teams of IT specialists and engineers worked on developing artificial intelligence models capable of automatically detecting explosive ordnance (EO) in images captured by drones.

The event took place at the Unit.City innovation park and was a continuation of the AI Labeling Sprint held earlier on 5 May in Kyiv. The goal of these hackathons is to implement innovative technologies in demining to more quickly clear Ukrainian territories contaminated with mines and EO.

Using over 8,000 training and 23,000 testing images, including aerial photographs annotated by participants of the first hackathon, participants of the second stage – AI Data Jam – developed highly effective AI algorithms over two days for the automatic identification of mines and other EO.

Participants worked both in-person and remotely, with mentorship provided by experts from organisations such as UADamage, The HALO Trust, and Dropla Tech.

The winning team, MineWatch AI, trained the most effective AI model and received a USD 2,000 certificate from UNDP in Ukraine and the Government of Luxembourg. They also gained the opportunity to further develop their product in collaboration with demining practitioners.

“What began as an experiment has grown into a full-fledged project with international support and tangible results. I thank everyone who dedicated their time and effort to this vital cause. The motivation is immense: together, we aim to create a unique product that will help clear Ukraine of mine contamination, making it safer metre by metre, without risking lives and with maximum use of technology,” said Deputy Minister of Economy, Environment and Agriculture of Ukraine, Ihor Bezkaravainyi.

Ben Lark, UNDP Ukraine Mine Action Programme Manager, stated during the event that Ukraine was demonstrating an impressive level of innovation in humanitarian demining. He remarked that what was being witnessed was the future being created in real time and emphasized that their goal was to make these innovations accessible to everyone working in the field. He added that for UNDP, it was significant that the hackathon brought together all key stakeholders, including government representatives, veterans, tech teams, civil society, and international partners.

The data collected and AI models developed during the hackathons will serve as the foundation for further training and refinement of AI algorithms. After testing, these technologies will be gradually integrated into humanitarian demining processes, significantly accelerating the clearance of Ukrainian land from mines and other explosive ordnance, making demining more efficient and safer. The solutions developed have the potential for application not only in Ukraine but also internationally.

The hackathon was organised by the Ministry of Economy, Environment and Agriculture of Ukraine, the Demine Ukraine initiative, the Centre for Humanitarian Demining, and the Ukrainian company UADamage, with support from the United Nations Development Programme (UNDP) in Ukraine and funding from the Government of Luxembourg.