Hierarchical Classification
-
Hierarchical active learning
[slides]
By:颜逸凡 2017-11-30
Experiments on three datasets show that the performance may be improved when we discard some labels which have few relevant instances. And we take different costs of query into account. The results are not consistent in three datasets. -
BatchRank: A Novel Batch Mode Active Learning Framework for Hierarchical Classification
[slides]
[paper]
By:颜逸凡
In many applications, the set of class labels are organized in a hierarchical tree structure, with the leaf nodes as outputs and the internal nodes as clusters of outputs. This paper proposed a novel batch active selection criterion based on the label structure.
Self-Paced Learning
-
Active self-pased learning: experiment
[slides]
By:唐英鹏 2017-11-20
-
Active self-pased learning: experiment
[slides]
By:唐英鹏 2017-11-13
-
公式推导
[slides]
By:唐英鹏 2017-10-20
A trial to find some common points between SPL abd AL by formula derivation -
公式合并
[slides]
By:唐英鹏 2017-10-13
Combined the formulation of SPL and Querying Discriminative and Representative Samples for Batch Mode Active Learning. -
Querying Discriminative and Representative Samples for Batch Mode Active Learning
[slides]
[paper]
By:唐英鹏 2017-09-29
This paper derive a novel form of upper bound for the true risk in the a