Abstract | ||
---|---|---|
We propose a generative Bayesian model that predicts instance labels from weak (bag-level) supervision. We solve this problem by simultaneously modeling class distributions by Gaussian mixture models and inferring the class labels of positive bag instances that satisfy the multiple instance constraints. We employ Dirichlet process priors on mixture weights to automate model selection, and efficiently infer model parameters and positive bag instances by a constrained variational Bayes procedure. Our method improves on the state-of-the-art of instance classification from weak supervision on 20 benchmark text categorization data sets and one histopathology cancer diagnosis data set. |
Year | Venue | Field |
---|---|---|
2014 | UNCERTAINTY IN ARTIFICIAL INTELLIGENCE | Data set,Dirichlet process,Instance-based learning,Bayesian inference,Pattern recognition,Computer science,Model selection,Artificial intelligence,Prior probability,Mixture model,Machine learning,Bayes' theorem |
DocType | Citations | PageRank |
Conference | 5 | 0.43 |
References | Authors | |
18 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Melih Kandemir | 1 | 182 | 16.91 |
Fred A. Hamprecht | 2 | 962 | 76.24 |