Towards Semi-Supervised Learning for Deep Semantic Role Labeling
This addresses the challenge of SRL for low-resource languages or domains by reducing dependency on extensive labeled corpora, though it is incremental as it builds on existing neural architectures.
The paper tackles the problem of neural models requiring large labeled datasets for semantic role labeling (SRL) by proposing a semi-supervised method that enforces syntactic constraints and uses unlabeled data, achieving F1 score improvements of +1.58 and +0.78 on limited labeled data (1% and 10%) over pre-trained models.
Neural models have shown several state-of-the-art performances on Semantic Role Labeling (SRL). However, the neural models require an immense amount of semantic-role corpora and are thus not well suited for low-resource languages or domains. The paper proposes a semi-supervised semantic role labeling method that outperforms the state-of-the-art in limited SRL training corpora. The method is based on explicitly enforcing syntactic constraints by augmenting the training objective with a syntactic-inconsistency loss component and uses SRL-unlabeled instances to train a joint-objective LSTM. On CoNLL-2012 English section, the proposed semi-supervised training with 1%, 10% SRL-labeled data and varying amounts of SRL-unlabeled data achieves +1.58, +0.78 F1, respectively, over the pre-trained models that were trained on SOTA architecture with ELMo on the same SRL-labeled data. Additionally, by using the syntactic-inconsistency loss on inference time, the proposed model achieves +3.67, +2.1 F1 over pre-trained model on 1%, 10% SRL-labeled data, respectively.