The Mathematical Corporation by Josh Sullivan & Angela Zutavern

Author:Josh Sullivan & Angela Zutavern
Language: eng
Format: epub
Publisher: PublicAffairs
Published: 2017-06-06T04:00:00+00:00

IMPOSSIBLE SOLUTIONS

In the third phase of the experimentation cycle, the challenge is devising ways to tackle the opportunity. How are you going to realize that vision? How are you going to create new and vital action within your new frame? This is not usually a straightforward process but a process that is iterative and experimental in nature. As you start brainstorming solutions, you may recognize you have the wrong frame and go back to the previous step.

A structured process to trigger fresh ideas generates the best strategies. Some of these ideas will come from your domain knowledge about the industry, your field of expertise, operations, and traditional data sets. Others will come from mash-ups of developments you see in other industries that you can apply to yours, just what John Brownstein did at UberHEALTH and Tadgh Smith did at ICE. But the really impossible solutions might surface from the results of implementing the techniques, capabilities, and potential of the mathematical corporation.

Ideas may evolve from playing out a new collaborative relationship between people and computers. Because the machine performs certain cognitive tasks better than people, you can gain potential by off-loading more tasks of comprehension, expression, perception, computation, memory, organization, and reasoning to the machine. What if you have more time for exercising your unique capabilities to imagine, create, and structure problem solving?

One development to consider is the potential for tapping other people’s brains through modern networks and social techniques, in particular, crowdsourcing. What if you throw the problem to people in other organizations or countries? If the problem is tightly framed, you can even work with an organization like Kaggle, the start-up that sponsors competitions to solve data science problems. If you work with Kaggle, you team up with a network of more than a half million data scientists.

In 2016, through Kaggle, State Farm attracted 517 data scientist teams to write an algorithm in which data from computer vision could be used to spot distracted drivers. Also through Kaggle, Expedia attracted 490 teams to predict which type of hotel customers would book. Santander Bank attracted 5,236 teams to predict satisfied and dissatisfied banking customers. Prizes for winning these competitions ranged from $25,000 to $65,000—a small expenditure to create a winning strategy through access to novel expertise.

Another means of brainstorming is a hackathon, a gathering of data scientists, coders, strategists, and others for an intensive day or two of prototyping. We worked with the Holocaust Museum’s Center for the Prevention of Genocide to host a hackathon to find a way to predict—and alert authorities to prevent—the next mass killing, wherever on the planet it might happen. Instead of assigning just a few experts to the project, we brought together a crowd to come at it from dozens or hundreds of angles—and ended up creating an algorithm that proves as reliable as the opinions of lifelong experts in genocide. (We come back to this example in Chapter 8.)

Even if you want to keep your data proprietary, you can benefit from math and algorithms produced elsewhere.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.