Multi-Step Inference for Reasoning Over Paragraphs
This work addresses the challenge of reasoning over paragraphs for natural language processing applications, representing an incremental improvement over existing symbolic and transformer-based methods.
The paper tackles the problem of complex reasoning over text by introducing a compositional model that combines neural modules for chained logical reasoning, achieving up to 29% relative error reduction on the ROPES dataset.
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.