International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
Automatic Question & Answer Generation from Paragraph Ms.Antara Dessai1 1Student, Department of Information Technology and Engineering, Goa College of Engineering, Farmagudi, Goa,
India ---------------------------------------------------------------------***--------------------------------------------------------------------It is motivated by the growing need for intelligent systems Abstract - In this paper, we present a novel approach that can comprehend textual information in a manner to question and answer generation utilizing the alike to human understanding. Traditional approaches to Bidirectional and Auto-Regressive Transformers Q&A generation often fall short in capturing nuanced (BART) model. Our method harnesses the power of contextual relationships within a given paragraph. BART's pre-trained representations to efficiently generate relevant questions from given contexts and In contrast, BART, a pre-trained transformer-based model, produce coherent answers. By fine-tuning BART on excels in capturing bidirectional dependencies, allowing it question-answer pairs, we achieve state-of-the-art to understand the intricate interplay of words and performance in generating natural and informative phrases. questions and their corresponding answers. We The primary goal of this project is to design, implement, demonstrate the effectiveness of our approach through and evaluate a Q&A generation system that leverages the extensive experimentation on various benchmark capabilities of BART. This involves fine-tuning the model datasets, showcasing its capability to generate diverse, on a carefully curated dataset, encompassing diverse contextually relevant questions and answers across topics and linguistic styles. The dataset comprises different domains. Furthermore, we explore the paragraphs accompanied by human-generated questions potential applications of our model in educational and corresponding answers, facilitating supervised platforms, conversational agents, and information learning for the model. retrieval systems. Overall, our work highlights the utility and versatility of the BART model in automating 1.1 Objectives question and answer generation tasks with high accuracy and fluency.This research presents a survey of Objective questions, the primary objective is to generate methods that could be used for objective and subjective question and answer pair generation from a given paragraph.
accurate questions that test factual knowledge or require specific answers. The BART model can analyze large datasets, identify key variables, and construct questions that cover a broad spectrum of topics with precision. By leveraging its ability to handle uncertainty and nonlinear relationships, the model can create diverse sets of objective questions that challenge learners and assess their understanding of the subject matter effectively.
Key Words: Question Answering (QA), Question Generation (QG), Bidirectional and Auto-Regressive Transformers (BART), Natural Language Toolkit (NLTK).
1.INTRODUCTION
For subjective questions, the goal shifts towards generating prompts that elicit thoughtful responses or opinions from individuals. Here, the BART model's strength lies in its capability to understand and represent complex human language nuances. It can generate prompts that stimulate critical thinking, creativity, or emotional expression, fostering deeper engagement and reflection. By incorporating diverse perspectives and linguistic variations, the model can produce subjective questions that cater to a wide range of preferences and experiences.
In recent years, advancements in natural language processing (NLP) have propelled the development of sophisticated models capable of understanding and generating human-like text. Among these, the task of Question and Answer (Q&A) generation from paragraphs has emerged as a pivotal application, with potential implications for education, information retrieval, and conversational agents. This project endeavors to explore and harness the power of state-of-the-art models, focusing on the Bidirectional and Auto-Regressive Transformers (BART), to achieve robust and context-aware Q&A systems.
© 2024, IRJET
|
Impact Factor value: 8.226
2. Related Work [1]. 1st Xu Chen & 2nd Jungang Xu proposed about the An Answer Driven Model For Paragraph-level Question
|
ISO 9001:2008 Certified Journal
|
Page 2391