Cold-Start Active Correlation Clustering
By: Linus Aronsson, Han Wu, Morteza Haghir Chehreghani
Potential Business Impact:
Finds groups of similar things by asking smart questions.
We study active correlation clustering where pairwise similarities are not provided upfront and must be queried in a cost-efficient manner through active learning. Specifically, we focus on the cold-start scenario, where no true initial pairwise similarities are available for active learning. To address this challenge, we propose a coverage-aware method that encourages diversity early in the process. We demonstrate the effectiveness of our approach through several synthetic and real-world experiments.
Similar Papers
Cold Start Active Preference Learning in Socio-Economic Domains
Machine Learning (CS)
Teaches computers to learn preferences from no data.
Data Skeleton Learning: Scalable Active Clustering with Sparse Graph Structures
Machine Learning (CS)
Makes computers group data better with less help.
Offline Clustering of Preference Learning with Active-data Augmentation
Machine Learning (CS)
Learns what people like from limited past choices.