Weakly Supervised Reasoning by Neuro-Symbolic Approaches
arXiv:2309.13072 · doi:10.3233/FAIA230162
Abstract
Deep learning has largely improved the performance of various natural language processing (NLP) tasks. However, most deep learning models are black-box machinery, and lack explicit interpretation. In this chapter, we will introduce our recent progress on neuro-symbolic approaches to NLP, which combines different schools of AI, namely, symbolism and connectionism. Generally, we will design a neural system with symbolic latent structures for an NLP task, and apply reinforcement learning or its relaxation to perform weakly supervised reasoning in the downstream task. Our framework has been successfully applied to various tasks, including table query reasoning, syntactic structure reasoning, information extraction reasoning, and rule reasoning. For each application, we will introduce the background, our approach, and experimental results.
arXiv commentCompendium of Neurosymbolic Artificial Intelligence, 665--692, 2023, IOS Press
References in corpus (11)
- Neural Machine Translation by Jointly Learning to Align and Translate
- Categorical Reparameterization with Gumbel-Softmax
- Character-level Convolutional Networks for Text Classification
- Axiomatic Attribution for Deep Networks
- Thumbs up? Sentiment Classification using Machine Learning Techniques
- Bidirectional Attention Flow for Machine Comprehension
- Attention is not Explanation
- Neural Paraphrase Generation with Stacked Residual LSTM Networks
- Learning to Paraphrase: An Unsupervised Approach Using Multiple-Sequence Alignment
- Jointly Learning Sentence Embeddings and Syntax with Unsupervised Tree-LSTMs
- Weakly Supervised Explainable Phrasal Reasoning with Neural Fuzzy Logic