U.S. Department of Energy

Pacific Northwest National Laboratory

CHISSL: A Human-Machine Collaboration Space for Unsupervised Learning

Publish Date: 
Thursday, June 1, 2017
We developed CHISSL, a human-machine interface that utilizes supervised machine learning in an unsupervised context to help the user group unlabeled instances by her own mental model. The user primarily interacts via correction (moving a misplaced instance into its correct group) or confirmation (accepting that an instance is placed in its correct group). Concurrent with the user's interactions, CHISSL trains a classification model on the user's grouping of the data. It then predicts the group of unlabeled instances, and arranging some of these alongside the instances manually organized by the user. We hypothesize that this mode of human and machine collaboration is more effective than active learning, wherein the machine decides for itself which instances should be labeled by the user. We found supporting evidence for this hypothesis in a pilot study where we applied CHISSL to organize a collection of handwritten digits.
Arendt DL, C Komurlu, and LM Blaha. 2016. "CHISSL: A Human-Machine Collaboration Space for Unsupervised Learning." In ACM CHI 2017. PNNL-SA-121349, Pacific Northwest National Laboratory, Richland, WA.
| Pacific Northwest National Laboratory