Wild Card Entrants Join IBM Watson AI XPRIZE Competition

May 08, 2018      

Today, the IBM Watson AI XPRIZE announced three additional teams for its multiyear contest, which is intended to encourage development of artificial intelligence tools to “solve societal grand challenges.” The contestants from the Wild Card round join 59 teams that have advanced to Round 2.

The IBM Watson AI XPRIZE is different from other competitions in that the participants identify the problems they want to solve rather than the organizers.

“The idea really is that we have an open competition — we’re not asking people to solve a given problem,” said Amir Banifatemi, new frontiers group lead at XPRIZE. “We’ve given the crowd out there a way to frame grand challenges with AI.”

“There are plenty of hackathons, as well as DARPA or NSF challenges out there that are domain-focused,” he said. “They’re all important, but the goal of this XPRIZE is to invite the community to identify problems.”

“It’s interesting to observe the emergence of interest around certain topics. It’s not top down,” Banifatemi said. “We’re trying to capture the essence of problem-solving approaches.”

IBM Watson AI XPRIZE logo 2

The Wild Card rounds are a way for the IBM Watson AI XPRIZE to include the latest research in rapidly evolving technologies.

“People need time [to create solutions], but AI is moving at such a rapid pace that teams and approaches we didn’t have last year or that we may not know about for next year need to be included,” Banifatemi told Robotics Business Review.

Drawing the AI Wild Card

“The Wild Card call opened up last year, and we got applicants over four months in the fall,” explained Banifatemi. “Our judges looked at 16 applicants, but 100 were interested, and three teams actually made it.”

Amir Banifatemi, new frontiers group lead for the IBM Watson AI XPRIZE

Letting the community define the societal problems to solve is a unique approach, says Amir Banifatemi, new frontiers group lead for the IBM Watson AI XPRIZE.

Netherlands-based OPTOSS AI is working on a platform that combines data from ground stations and satellites with crowdsourced data to improve disaster notifications. It would compare past weather events and use predictive analysis to help emergency response organizations apply resources where they’re needed most.

“There are a lot of humanitarian organizations that need to send people on the ground, such as Doctors Without Frontiers,” Banifatemi said. “How can AI help with the best zones for sending and monitoring aid, water, crops, and animals?”

U.S.-based Mt. Cleverest plans to combine natural language processing and crowdsourced content to develop a free, “self-improving” online textbook to “to improve education outcomes at a global scale,” according to the nonprofit XPRIZE organization.

It also hopes to create customized curricula, provide teachers with automated assistance, and be able to monitor student progress in real time.

“This is not another Wikipedia, but it is helping students help one another,” Banifatemi said. “Imagine learning being more widely dispersed and AI capturing educational mechanisms and patterns from the open-source community.”

Israel-based Zzapp Malaria plans to apply AI in the fight against malaria, which still infected 200 million people and killed more than 400,000 in Africa just last year. Zzapp Malaria has proposed using machine learning to analyze online databases and satellite data to create intervention strategies for individual communities based on environmental and other conditions.

“Zzapp Malaria isn’t the only group trying to solve the problem, but the use of AI to help against this disease hasn’t really been done yet,” said Banifatemi.

“These three Wild Card teams give incremental value to existing teams,” he added. “The judges determined that they have the staffs and plans to do what they need to.”

Local ambassadors and more milestones

“We have local ambassadors in London, New York, Paris, and Pittsburgh,” Banifatemi said of the XPRIZE organization. “We have many strong advocates. For instance, Pittsburgh had nine teams to start with connected to Uber, Google, and other companies.”

“It had meetups, and now six teams are participating,” he added. “There’s a huge effervescence created locally by ambassadors who understand this competition as a catalyst for dialogue and for solving problems. Events provide the companies with more exposure.”

“We also have crossovers with other prizes, like around the oceans or AI,” Banifatemi noted. “There other efforts we undertake that aren’t visible to the public.”

The IBM Watson AI XPRIZE teams must provide an annual report on their progress toward certain milestones in August. They also have to be cognizant of regulatory requirements.

“All teams must demonstrate certain things — security, data privacy, human-machine interaction, and collaboration,” said Banifatemi. “We’ll also provide judging criteria complementary to the milestones, for example, around machine vision and learning for an educational project.”

“We’ll have internal review boards for privacy, but healthcare teams also have to be HIPPA-compliant, while manufacturing or logistics need to have OSHA compliance,” he added. “Teams will have a chance to apply for the yearly milestone prize, plus the top 10, which provide a small monetary prize to help out.”

More Wild Card teams to come

“The last opportunity for Wild Card teams to join will be between August and October or November,” Banifatemi said. “They’ll then have 18 months to prove themselves.”

Later contestants will have some catching up to do, but they’ll be judged by the same criteria as other IBM Watson AI XPRIZE participants.

“Knowing what others have done should put them on par,” said Banifatemi. “All teams know the competition timeline — everything is transparent. We’ve tried to be open to new ideas, and it’s up to the teams to prove that they are capable.”

“We’re not limiting the number of teams joining, but they have to read the guidelines and prepare applications,” he noted. “Not many people understand AI, and we’re being pragmatic, but there’s no limit.”

The XPRIZE organizers announced the top 10 competitors and top 2 teams so far, but there is a lot of work yet to be done. In addition, judges and advisers may be added as more domain expertise is needed.

“We have 33 judges and about 50 experts, and we can add more organically,” Banifatemi said. “For example, health is a big topic among the teams, and there’s a full ecosystem of experts around public health.”

Long-term benefits

As with any grand challenge, the IBM Watson AI XPRIZE could result in an array of long-term benefits, even from teams that aren’t ultimately winners.

“Regardless of the grand prize winner [to be announced in April 2020], all of them are creating sparks,” Banifatemi said. “Some are startups, some are doing research in labs, and some are teams of individuals from corporations.”

“Whether they get funded or drop out [of the competition], all of these teams are doing interesting things,” he said. “They could provide inspiration to others working to solve similar problems.”