Gamified citizen science, teamed with the potential of AI, could help power significant advances in cancer research, a scientist has suggested.
UC Merced cognitive science professor Jeff Yoshimi came up with the idea when his wife learned she had breast cancer.
Toshimi wondered if the battle against cancer could be scaled up with a suite of video games. From that first flash in 2013 and over 11 years, he expanded the idea into a compelling action plan and put it in a book. “Gaming Cancer,” published by MIT Press.
“Much of this book was written in cancer wards and chemo rooms,” Yoshimi says in the first chapter.
“Scientific problems associated with cancer are messy and complex,” he wrote, “but in some cases they have core logic that can be presented in a game context with its simplified rules.”
Citizen science games are out there, though on a far smaller scale than what Yoshimi lays out in the book.
One example: “Foldit” makes a video game out of how proteins restructure themselves. Outcomes from this “protein folding” game can help scientists understand cellular mutation and runaway replication of diseased tissue. It can help perfect therapeutic drugs.
In 2020, about 750,000 users played “Foldit.” An impressive number, but Yoshimi is thinking bigger. Like “Baldur’s Gate” bigger. “Candy Crush” bigger.
Yoshimi has hands-on experience with game-style programs. In 2008 he created Simbrain, where users design neural networks with simplified visuals. He uses Simbrain in his courses and research, helping students and colleagues grasp aspects of neuroscience, psychology and AI.
In an interview, Yoshimi, a UC Merced founding faculty member who came aboard in 2004, talked about the knowledge that could be unleashed by scaling up citizen science. He told a story about going to a restaurant with his father and seeing a boy whose knee bounced like a jackhammer.
“My dad said, ‘I wonder if we could make a little generator, attach it to that knee and capture all that energy,’” Yoshimi said.
“That stuck with me. People have these exquisitely evolved brains and we enjoy solving puzzles. It’s good for the species. So we have this huge reservoir of untapped problem-solving power, this Earth-wide computing effort being used on ‘Grand Theft Auto.’
“What if we redirected some of that to problems of broad significance, like cancer?”
Yoshimi discusses the “meta-game” — the process of creating this suite of powerful, irresistible games. He details everything they need: crack programmers, visionary leaders and deep-pocketed financiers, as well as reward structures for users and smooth links between gamers and labs.
At the centre of it all Yoshimi envisions Simbody, a game engine that simulates biological systems from whole bodies to nanoscale.
Start with a cancer conundrum. Break down the problem until there’s one task a game could tackle. Plug it in. Yoshimi said the suite, which he dubbed “Cancer Wars,” could use all the game styles: action, adventure, role-playing, first-person shooter, strategy, sports.
Yoshimi said many of the existing citizen science games have AI elements that work hand-in-hand with the player, matching the former’s data-processing power with the latter’s power of intuition.
“We want to have human-AI symbiosis,” he said. “AI can do raw number crunching and statistical generalization. Humans see the bigger picture, the relevance of one thing to another, the creative insight that a machine finds harder to capture.
“AI has made big jumps, but it’s not at a point where we push a button, ChatGPT thinks about it for a month and solves the problem. The best games will take AI as far as it can go, interweaved with human intelligence, so the experience is seamless.”
Yoshimi hopes “Gaming Cancer” offers a roadmap for ramping up video games in the fight against cancer. He added that eradicating the disease is a worthy goal but certainly not the only one. There are countless important victories within reach. Detection and treatments for the hundreds of cancer types demand attention. Games can foster an understanding of clinical research and promote lifestyle choices that can reduce cancer risk.
“It’s worth the effort,” Yoshimi said. “Even if you don’t hit the moon shot, all the intermediate shots are valuable.”
