What can we learn from Semantic Tagging?
This work addresses performance enhancement in NLP tasks through multi-task learning, but it is incremental as it builds on existing methods with a new auxiliary task.
The paper investigated using semantic tagging as an auxiliary task in multi-task learning for NLP tasks like part-of-speech tagging, dependency parsing, and natural language inference, finding that a 'learning what to share' setting consistently improved performance across all tasks.
We investigate the effects of multi-task learning using the recently introduced task of semantic tagging. We employ semantic tagging as an auxiliary task for three different NLP tasks: part-of-speech tagging, Universal Dependency parsing, and Natural Language Inference. We compare full neural network sharing, partial neural network sharing, and what we term the learning what to share setting where negative transfer between tasks is less likely. Our findings show considerable improvements for all tasks, particularly in the learning what to share setting, which shows consistent gains across all tasks.