For thousands of years, people have yearned to know the future. They’ve peered into the entrails of animals, gone to the mountaintop oracle, planted with the Farmers Almanac, or bet with Nate Silver on basketball’s March madness. And for just as long people have been disappointed when the predictions failed.
Can new advances in science succeed in making predictions, well, more predictive?
The University of Vermont’s 2014 Macmillan Symposium, “Prediction—The Next Big Thing,” will tackle that question head-on.
Sponsored by UVM’s Complex Systems Center, the symposium will be held on Monday, April 28, from 1 to 4 p.m. at the Davis Center's Grand Maple Ballroom. The event is free and open to the public.
Two leading scholars working in the realm of complex social systems, Neil Johnson, a physicist at University of Miami, and César A. Hidalgo, from MIT’s Media Lab, will speak. Along with an experienced group of UVM faculty and students, they will explore today’s landscape of scientific prediction.
Futures forecast
“We have contended with — and discarded — the possibility of a deterministic, mechanistic universe, all paths laid out from the start,” writes Peter Dodds, director of UVM’s Complex Systems Center who is helping to lead the event.
“But as the physical sciences have grown, we have had much success in many areas: we have described the fundamental unpredictability of the quantum world,” Dodds notes, “and we have steadily improved our ability to predict the weather and certain natural disasters, crucially quantifying and explaining our uncertainty.”
Into this world of quantified uncertainty and post-Heisenberg complexity, a new generation of scholars of prediction have jumped — with both eyes open. Some of the speakers at the symposium will talk about what they’re discovering — from taming fibrillating hearts, to forecasting civil unrest, to building biological gene circuits.
Schedule
At 1 p.m., César Hidalgo, head of the Macro Connections group at the MIT Media Lab and the ABC Career Development Professor at MIT, will lead off a group of short talks. An “antidisciplinary academic by choice and a poet at heart,” Hidalgo’s work focuses on improving understanding of the world’s complexity. His tools include the “construction of visualization engines that make available unwieldy volumes of data,” and the development of new data collection methods, “that make visible hitherto neglected aspects of our reality.” He seeks, he says, “development of theories and narratives that can help put together the pieces that reductionist approaches have pulled apart.”
From 2 to 3 p.m. a variety of posters on prediction by students in UVM’s complex systems group will be presented — with a chance for attendees and presenters to talk together.
Then, at 3 p.m., Neil Johnson will speak, leading off the next group of short talks. Johnson heads up a new inter-disciplinary research group in complexity at the University of Miami looking at collective behavior and emergent properties in a wide range of real-world complex systems from medicine to finance.
Big data window
“Prediction runs from the mundane and individual — knowing it might rain today means we should bring an umbrella,” Dodds writes, “— to the potentially disastrous and widespread — an incoming category 5 hurricane leading to the evacuation of cities.”
Of course, much of the work of prediction we take for granted, and the predictions work every time. “All complex life employs algorithmic inferences about the future,” Dodds notes. In other words, when you jump, you expect to leave the ground and have a very good chance of coming back down again.
But for centuries, some thinkers have doubted the human capacity to predict much beyond that. “The future is doubtful,” wrote Seneca in 55BC. “Tomorrow is an old deceiver,” wrote Samuel Johnson. Even Albert Einstein himself, so deeply fascinated with what might be known, wrote, “I never think about the future.”
And yet biology would suggest that prediction is more achievable and powerful than this skeptical tradition might suggest. “Strong predictive capabilities are essential in the full spectrum of evolutionary systems,” Dodds notes — and from evolution many of the new insights into prediction are emerging.
In addition, “the advent of global, interconnected sociotechnical systems,” Dodds writes, “and their quantification in ‘Big Data’ would seem to hold much promise for our ability to greatly expand the scope of prediction science.” Soon, we may know if he’s right.
This event is underwritten by the MacMiillan Family Trust.