Explore a 21-minute conference talk from the ACM User Interface Software and Technology Symposium that introduces the concept of exploratory labeling for large document collections. Discover how computational and interactive methods can assist analysts in categorizing documents into unknown and evolving labels. Learn about a proposed interactive visual data analysis method that combines human-driven label ideation with machine-driven recommendations. Understand the process of progressive label discovery, ideation, and refinement using unsupervised machine learning techniques. Examine the evaluation of this method through a real-world labeling problem and controlled user studies, gaining insights into emerging interaction patterns in exploratory labeling activities.
The Exploratory Labeling Assistant - Mixed-Initiative Label Curation with Large Document Collections
Association for Computing Machinery (ACM) via YouTube
Overview
Syllabus
Introduction
Problem Statement
Property Labeling
User Profile
Example
Demo
Agari Example
User Study
User Focus
Discovery
Evaluation
Results
Conclusion
Taught by
ACM SIGCHI