Improving Out-of-Scope Detection in Intent Classification by Using Embeddings of the Word Graph Space of the Classes.
Paulo R. CavalinVictor Henrique Alves RibeiroAna Paula AppelClaudio S. PinhanezPublished in: EMNLP (1) (2020)
Keyphrases
- class labels
- equivalence classes
- low dimensional
- selecting a small number
- pattern recognition
- decision trees
- graph theory
- search space
- classification accuracy
- object detection
- supervised learning
- pattern classification
- classification method
- space time
- feature extraction
- discriminant functions
- multiple classes
- vector space
- classification algorithm
- support vector machine svm
- keywords
- feature set
- feature vectors
- class membership
- machine learning
- digital mammograms
- binary classifiers
- directed acyclic graph
- weighted graph
- directed graph
- false positives
- detection method
- text classification
- training set