Syntactic Scaffolds for Semantic Structures
arXiv:1808.10485
Abstract
We introduce the syntactic scaffold, an approach to incorporating syntactic information into semantic tasks. Syntactic scaffolds avoid expensive syntactic processing at runtime, only making use of a treebank during training, through a multitask objective. We improve over strong baselines on PropBank semantics, frame semantics, and coreference resolution, achieving competitive performance on all three tasks.
arXiv commentAccepted at EMNLP 2018
References in corpus (12)
- Adam: A Method for Stochastic Optimization
- Efficient Estimation of Word Representations in Vector Space
- Neural Machine Translation by Jointly Learning to Align and Translate
- Training Very Deep Networks
- SQuAD: 100,000+ Questions for Machine Comprehension of Text
- Learning to Map Sentences to Logical Form: Structured Classification with Probabilistic Categorial Grammars
- Segmental Recurrent Neural Networks
- Frame-Semantic Parsing with Softmax-Margin Segmental RNNs and a Syntactic Scaffold
- Learning Global Features for Coreference Resolution
- Adversarial Multi-Criteria Learning for Chinese Word Segmentation
- Deep Multitask Learning for Semantic Dependency Parsing
- Learning Joint Semantic Parsers from Disjoint Data