Multi-Step Inference for Reasoning Over Paragraphs
arXiv:2004.02995 · doi:10.18653/v1/2020.emnlp-main.245
Abstract
Complex reasoning over text requires understanding and chaining together free-form predicates and logical connectives. Prior work has largely tried to do this either symbolically or with black-box transformers. We present a middle ground between these two extremes: a compositional model reminiscent of neural module networks that can perform chained logical reasoning. This model first finds relevant sentences in the context and then chains them together using neural modules. Our model gives significant performance improvements (up to 29\% relative error reduction when comfibined with a reranker) on ROPES, a recently introduced complex reasoning dataset.
arXiv commentaccepted by EMNLP 2020
References in corpus (7)
- Learning to Retrieve Reasoning Paths over Wikipedia Graph for Question Answering
- Learning to Reason: End-to-End Module Networks for Visual Question Answering
- Exploring Graph-structured Passage Representation for Multi-hop Reading Comprehension with Graph Neural Networks
- Question Answering via Integer Programming over Semi-Structured Knowledge
- Multi-hop Question Answering via Reasoning Chains
- Neural Module Networks for Reasoning over Text
- Constituency Parsing with a Self-Attentive Encoder