Entity Guided Question Generation with Contextual Structure and Sequence Information Capturing

Abstract

Question generation is a challenging task and has attracted widespread attention in recent years. Although previous studies have made great progress, there are still two main shortcomings: First, previous work did not simultaneously capture the sequence information and structure information hidden in the context, which results in poor results of the generated questions. Second, the generated questions cannot be answered by the given context. To tackle these issues, we propose an entity guided question generation model with contextual structure information and sequence information capturing. We use a Graph Convolutional Network and a Bidirectional Long Short Term Memory Network to capture the structure information and sequence information of the context, simultaneously. In addition, to improve the answerability of the generated questions, we use an entity-guided approach to obtain question type from the answer, and jointly encode the answer and question type. Both automatic and manual metrics show that our model can generate comparable questions with state-of-the-art models. Our code is available at https://github.com/VISLANG-Lab/EGSS.

Automatic Extraction of Standard Multimodal Knowledge for the Petroleum Field

“…The text clustering uses the K-means method, which is one of the most used clustering methods with good clustering results and fast speed [4]. The document similarity calculation method adopts the vector space model.…”

Ask To The Point: Open-Domain Entity-Centric Question Generation

“…Question generation (QG) aims to automatically generate questions from raw texts , knowledge bases (Bi et al, 2020), or images (Vedd et al, 2022). For text-based QG, there are mainly two settings, answer-aware (Huang et al, 2021;Wu et al, 2022) and answer-agnostic (Back et al, 2021;Zhao et al, 2022). The difference is whether answers are given or not.…”

“…Particularly, text-based QG broadly benefits conversational chatbots to improve user interaction (Gao et al, 2019), educational materials to enhance reading comprehension , or QA dataset enrichment to boost QA development (Lyu et al, 2021). There are mainly two QG settings, answeraware (Huang et al, 2021;Wu et al, 2022) and answer-agnostic (Back et al, 2021;Zhao et al, 2022), the difference between which is whether answers are known or not.…”

Automatic Generation of Socratic Subquestions for Teaching Math Word Problems

“…Prior work on guided and controlled question generation uses either entities as guiding mechanism (Huang et al, 2021) or reinforcement learningbased graph to sequence approach . Identification of entities and relationships present in the text often uses rule-based or on-shelf extraction tools, which are hard to extend (Dhingra et al, 2020).…”