Leonardo Iheme has won the 2023 SAS Hackathon award for his Artificial Intelligence (AI) solution to the Turkey earthquake where about 50,000 people lost their lives and hundreds of thousands were injured in February (2023).
Iheme (Nigerian), James Swinnerton (UK-based) and Can Tosun all from EPAM System had entered the competition and won in two categories –the natural language and the regional awards (Europe, Middle East and Africa).
The February devastating earthquake in the Republic of Türkiye put the country in disarray. One of the problems faced by aid givers was inability to pinpoint the areas that needed the most help. Tweets and hashtags were popping up from different parts of the country that went viral.
Leonardo, popularly known as Leo, said:
“When people tweet about the number of people who need help in a particular area, we could aggregate those tweets but there was no infrastructure in place to aggregate those tweets.
That was when we saw the opportunity and decided to use natural language processing to gather all these from Twitter, convert them into addresses, convert these addresses into geolocation through longitude and latitude.
Natural Language processing helped a lot in processing these tweets. With these locations that were converted, we were able to get satellite images and we could compare ‘before and after’ the images for confirmation of the quake. For example, saying that there used to be 50 buildings here and now there are only 30 of them now because of the earthquake. We were able to gather such data and this really helped the aid workers to prioritise the most distressed scenes. That was compelling enough for the judges of the panel.”
Leonardo and his team had won a Turkish national award for developing an AI to detect tumours in breast cancer. The government organised the competition on using AI to solve health-related problems.
Leonardo Iheme further stated:
“It’s my passion as well, so I led a team to win this AI Competition for health.
We had a couple of the organisers shared ultrasound images – some with tumours and others without tumours and the goal is to accurately detect the tumours because that would help the doctors to localise it, that’s the challenge the radiologists have. Sometimes, they look at these images but can’t make decisions because they can’t see the tumour and if there is a tumour, how big it is. And this is where AI can really assist them.
We entered this competition and we emerged the winner. I was particularly proud of that because I led the data scientists there. We emerged the winner at that big event where the President of Turkey was in attendance and greeted us.”
