Abstract:Machine Reading Comprehension (MRC) is one of the most challenging tasks in both NLP and IR researches. Recently, a number of deep neural models have been successfully adopted to some simplified MRC task settings, whose performances were close to or even better than human beings. However, these models still have large performance gaps with human beings in more practical settings, such as MS MARCO and DuReader datasets. Although there are many works studying human reading behavior, the behavior patterns in complex reading comprehension scenarios remain under-investigated. We believe that a better understanding of how human reads and allocates their attention during reading comprehension processes can help improve the performance of MRC tasks. In this paper, we conduct a lab study to investigate human's reading behavior patterns during reading comprehension tasks, where 32 users are recruited to take 60 distinct tasks. By analyzing the collected eye-tracking data and answers from participants, we propose a two-stage reading behavior model, in which the first stage is to search for possible answer candidates and the second stage is to generate the final answer through a comparison and verification process. We also find that human's attention distribution is affected by both question-dependent factors (e.g., answer and soft matching signal with questions) and question-independent factors (e.g., position, IDF and Part-of-Speech tags of words). We extract features derived from the two-stage reading behavior model to predict human's attention signals during reading comprehension, which significantly improves performance in the MRC task. Findings in our work may bring insight into the understanding of human reading and information seeking processes, and help the machine to better meet users' information needs.

A Prompt-based Few-shot Machine Reading Comprehension Model for Intelligent Bridge Management

CPCM-MRC: A Few-shot Machine Reading Comprehension Approach for Intelligent Bridge Management via Continual Pre-training and Copy Mechanism

Bridge Relation Extraction: New Chinese Dataset and Model

Bridge inspection named entity recognition via BERT and lexicon augmented machine reading comprehension neural model

Feeding What You Need by Understanding What You Learned

A Multi-answer Multi-task Framework for Real-world Machine Reading Comprehension.

Hierarchical Answer Selection Framework For Multi-Passage Machine Reading Comprehension

Towards Medical Machine Reading Comprehension with Structural Knowledge and Plain Text.

Frame-based Neural Network for Machine Reading Comprehension

A Study of the Tasks and Models in Machine Reading Comprehension

Towards Human-level Machine Reading Comprehension: Reasoning and Inference with Multiple Strategies.

Intelligent Question Answering System Based on Machine Reading Comprehension

Reinforced Mnemonic Reader for Machine Reading Comprehension

Read, Attend, and Exclude: Multi-Choice Reading Comprehension by Mimicking Human Reasoning Process

Human Behavior Inspired Machine Reading Comprehension

Reinforced Mnemonic Reader for Machine Comprehension

Retrospective Reader for Machine Reading Comprehension

Multi-Passage Machine Reading Comprehension with Cross-Passage Answer Verification.

MPrompt: Exploring Multi-level Prompt Tuning for Machine Reading Comprehension

Integrate Candidate Answer Extraction with Re-Ranking for Chinese Machine Reading Comprehension

Chinese Machine Reading Comprehension Based on Language Model Containing Knowledge.