STS-UHH at SemEval-2017 Task 1: Scoring Semantic Textual Similarity Using Supervised and Unsupervised Ensemble.
Sarah KohailAmr Rekaby SalamaChris BiemannPublished in: SemEval@ACL (2017)
Keyphrases
- word sense induction
- semantic similarity
- word sense disambiguation
- supervised learning
- unsupervised learning
- supervised methods
- word sense
- semi supervised
- natural language
- unsupervised methods
- learning algorithm
- sense disambiguation
- similarity measure
- semantic similarity measure
- sentence similarity
- feature selection
- supervised classification
- semantic information
- weakly supervised
- training data
- unsupervised feature selection
- ensemble learning
- training set
- discriminant projection
- supervised training
- neural network
- pointwise mutual information
- high level
- distance measure
- low level features
- distance function
- unlabeled data
- semantic web
- active learning
- ensemble methods
- decision trees
- multimedia
- metadata
- text categorization
- semantic labels
- keywords
- pairwise
- feature space