Israeli AI Firm Fights Malaria With Machine Learning

Dozens of teams are competing to apply machine learning and artificial intelligence to solving serious social problems in the IBM Watson AI XPRIZE. A relatively late entry is Zzapp Malaria, which is working to eradicate one of humanity’s most persistent diseases. The company was recently named as one of the current top 10 teams in the multiyear XPRIZE competition, at the NeuIPS conference.

In 2017, more than 219 million people were infected with malaria, and more than 435,000 people in Africa died from the mosquito-borne disease. Pregnant women and children under five were most of the fatalities.

Israel-based Zzapp Malaria is working to use AI to analyze satellite data and online databases and optimize localized responses such as spraying pesticides.

Robotics Business Review recently spoke with Zzapp Malaria co-founders Arnon Houri Yafin and Ari Eichler about their efforts. Houri-Yafin is also CEO of the organization, which joined the IBM Watson AI XPRIZE contest as a “wild card” team.

Precision required in fighting disease

“Malaria can be eliminated, and it has been in many countries,” said Houri Yafin. “In order to do so, operations have to be perfect, both in terms of not missing any house or water body, as well as in terms of tailoring the intervention strategy to the particular village or town, the spraying techniques and materials, the season, the duration, etc.”

“Malaria could be eliminated, but you’d need a perfect operation,” Houri-Yafin said. “You can’t miss any water bodies [with pesticides].”

“Zzapp Malaria customizes solutions to villages,” said Eichler. “We decided to develop a system to help workers find every water body, to aid in decision-making on where to apply and when.”

Zzapp Malaria builds AI, app

“Our product has two components: a mobile app and a planning tool. The app ensures things actually happen on the ground,” explained Houri Yafin. “We have conducted two pilots in Ghana, actual controlled trials, and demonstrated that the app makes a big difference.”

“We are now working on turning the prototype into a product,” he said. “The AI component requires more work — especially feeding more data into our algorithm. We are trying to establish collaborations with countries that are fighting malaria, as well as to run our own operations with a variety of malaria-control methods.”

“We are launching our first version in three months,” said Houri Yafin. “We see our system as the Waze of malaria elimination. There’s an element that gives you directions so you don’t make mistakes — ‘Turn right here; spray this house.’”

“Then there’s the element of optimization — finding the fastest, most efficient route,” he added. “The thing is that both of these elements depend on each other. I couldn’t communicate the route without the component of directions. By the way, Nitzan Ron, our user interface designer, used to work at Waze.”

“We’re very close to the point where the two points would be connected with each other and the customer,” said Houri Yafin.

Joining the IBM Watson AI XPRIZE entries

“We heard about XPRIZE from a friend,” said Houri Yafin. “He read about the ‘Wild Card’ round and thought it was just right for Zzapp, since we are using AI to generate a significant impact on the world.”

“The core team is [Chief Operating Officer] Abel Vigodny, Ari, and me,” he added. “We just have a lot of people helping us — someone helping us design a neural network, and three developers for the mobile app and GUI of Zzapp.”

Houri Yafin said the group was proud to be recognized as a top 10 team so soon after joining the competition. “The XPRIZE competition is not just about buzzwords and trying to make an impression,” he said. “They actually look at the impact that the groups are trying to make and whether the technology is suitable for that goal. When we heard about our results, we were very excited. We are proud to be recognized.”

“We’re now taking the prototype app and making it into a product,” he added. “One challenge is with data collection. We’re trying to solve this in two ways: We’ll collaborate with any country that wants to use data to train the algorithm, as well as do more operations with us in the field to collect more data and get more intervention in context.”

The IBM Watson AI XPRIZE judges provide detailed feedback, and the international organization makes experts available for mentoring.

“We haven’t yet used XPRIZE for guidance, but we’re certainly counting on it in the future,” said Houri Yafin.

More goals to meet

“In the coming year, in addition to our own goals, we’ll have some objectives for all teams to accomplish by October [2019] to be judged as a top 10 team,” said Amir Banifatemi, general manager for innovation and growth at XPRIZE. “There are goals for human-machine collaboration, ethics, and transferability and applicability.”

“Despite the wide initial paths, there will be two technical tests,” he added. “We’ll have an all-encompassing set of tests, suites, and environments. We’ll also have more judges, who’ll be getting more teams ready in two ‘boot camps.’”

XPRIZE is also working on adding 5G, artificial intelligence, and virtual reality to physical robots through the ANA Avatar prize. The goal is to use robotics to project a human’s skills and participation somewhere else.

(Editor’s note: This is part two of a two-part profile of some of the top teams in the $5 million IBM Watson XPRIZE competition. Read part one here.).