Dynamic Fusion Networks for Machine Reading Comprehension

TL;DR

This paper introduces Dynamic Fusion Networks, a neural model for machine reading comprehension that dynamically adapts its attention and reasoning strategies for each question, achieving state-of-the-art results on the RACE dataset.

Contribution

The paper proposes a novel neural architecture with dynamic multi-strategy attention and reasoning, optimized via reinforcement learning for improved MRC performance.

Findings

Achieves best results on RACE dataset

Produces more effective attention vectors than other models

Demonstrates flexible, sample-specific network architecture

Abstract

This paper presents a novel neural model - Dynamic Fusion Network (DFN), for machine reading comprehension (MRC). DFNs differ from most state-of-the-art models in their use of a dynamic multi-strategy attention process, in which passages, questions and answer candidates are jointly fused into attention vectors, along with a dynamic multi-step reasoning module for generating answers. With the use of reinforcement learning, for each input sample that consists of a question, a passage and a list of candidate answers, an instance of DFN with a sample-specific network architecture can be dynamically constructed by determining what attention strategy to apply and how many reasoning steps to take. Experiments show that DFNs achieve the best result reported on RACE, a challenging MRC dataset that contains real human reading questions in a wide variety of types. A detailed empirical analysis also demonstrates that DFNs can produce attention vectors that summarize information from questions, passages and answer candidates more effectively than other popular MRC models.