Escape from Wonderland

Implicit Crowdsourcing Design

Most projects start with user-centered research methods such as interviewing and competitive analysis to find a need and then design a solution to solve it. However, when designing a solution for human computation problems, we had to find a problem that computers had difficulty solving that human computation could solve. For Interaction Design Studio, our group designed a mobile game that players could compete for high scores and solve challenges that took their responses and used them to train machine learning algorithms. I was responsible for designing the game world, but I also worked collaboratively on concept generation, game mechanics, and interaction design.

My Role

Research & Design

Team Members

Jonathan Brown
Gaby Moreno
Anthony Wong

Concept Generation

Escape from Wonderland was inspired by the work of Louis von Ahn, the creator of reCAPTCHA and more recently, Duolingo. Both of these products use implicit crowdsourcing, marrying a free service that the user wants with a task that they are able to perform better than computers, the output of which a company is willing to pay for. We brainstormed several different tasks such as using foreign students to translate menus, mapping areas unreachable by cars such as college campuses, and even judging public sentiment at Olympic games. We found that deciphering emotions in human language was an area that computers still had trouble understanding, especially if a person was being sarcastic. We found that machine learning software could only learn about tones if a large number of audio clips are properly tagged by people who can understand tones such as sarcasm.
whiteboard

Sketches and Wireframes

We wanted to create a game world similar to Temple Run where players could swipe in order to dodge obstacles and collect cards for points.
sketches

Iterations

The design went through numerous iterations to create a whimsical yet realistic game world.
alice_iteration_screens