ENHANCING ARABIC TEXT COMPREHENSION THROUGH Automated Grammar-Aware Question Generation By: Deema Jabi , Raneen Sawalmeh UNDER THE SUPERVISION OF DR. AMJAD HAWASH INTRODUCTION Improving the understanding and automated analysis of Arabic texts has become essential in rapid digital progress, particularly given the rising need for intelligent systems communicating with natural language. This project focuses on enhancing Arabic text comprehension through the development of an automated, grammar-aware question generation system. Problem Statement 1- Complex morphology and syntax 3-Lack of precise educational tools 2- Limited availability of AI systems for Arabic question generation and challenges in grammar-aware formulation. 01. Develop an automated Arabic text comprehension system using advanced NLP models Proposed Solution 02. Generate contextually relevant and grammatically accurate questions 03. Integrate AI-based assessments 04. Developing Tutorial Modules 01. 02. Automated Grammar-Aware Question Generation 03. Project Objectives Enhance Arabic Text Comprehension Enhance the Educational Process Using AI RELATED WORKS Summary of Previous Works: Focused on Arabic text processing using NLP, particularly syntactic and semantic analysis. Differences Between the Current Project and Previous Works This project differentiates itself by integrating grammar-aware strategies to automatically generate contextually accurate questions, addressing the complexities of Arabic grammar in a novel way. 01. 02. Processing Layer 03. NLP Layer System Architecture Presentation Layer 01. Stanza System Activity 02. google ai 03. gTTS Tools and Libraries Used: A natural language processing library that supports tasks like syntactic analysis and named entity recognition. A library that converts text to speech using Google's Text-to-Speech API . A Python library for natural language processing, offering simple tools like tokenization, stemming, and text analysis. Integrated Development Environment Jupyter Notebook Google Colab 01 02 03 04 Algorithms 1-Word Classification Based on Syntactic Context Break the sentence into individual words. WORD SEGMENTATION: Classify each word based on its syntactic type (noun, verb, adjective, pronoun, etc.). BASIC CLASSIFICATION: Examine the neighboring words (before and after) to understand the context. CONTEXTUAL ANALYSIS: Identify named entities such as people, places, or organizations (e.g., "Ahmed" or "Cairo"). NAMED ENTITY RECOGNITION: 01 02 03 04 Algorithms 2- Create questions using input text Setting up the work environment By installing a library google.generativeai as genai SET UP THE ENVIRONMENT Configure the connection to the Google Gemini API CONNECT TO THE SERVICE Choose the text that will be used to generate the questions. DEFINE THE INPUT TEXT Request question generation from the input text via the API. SEND THE TEXT TO THE SERVICE 05 Get the generated questions. RECEIVE THE QUESTIONS TEST & TRAIN 1️⃣Dataset Preparation: Collected diverse text samples from multiple domains to test system robustness From different sources. 2️⃣ Preprocessing Evaluation: Assessed the performance of keyword extraction, lemmatization, and stop-word removal. 3️⃣ Recommendation Accuracy: Compared AI-generated questions with human-generated ones using Precision, Recall, and F1-score. 4️⃣ User Engagement: Collected feedback through user surveys on usefulness, relevance, and clarity of generated questions. TEST & TRAIN 5️⃣ Real-Time Performance: Measured system response time for different input lengths and complexity levels. 6️⃣ Personalization Evaluation: Assessed the system’s ability to generate context-aware questions based on input variations. 7️⃣ Scalability Tests: Evaluated system performance with large-scale text inputs. 8️⃣ Error Handling: Tested system resilience against incomplete or erroneous inputs. CONCLUSION AI-powered system enhances comprehension of Arabic texts through grammar-aware question generation. NLP techniques ensure contextually accurate questions. The Tkinter-based GUI makes the system user-friendly and interactive. The integration of text-to-speech enhances learning and engagement. FUTURE WORK Improving Language Analysis Algorithms Supporting Additional Dialects and Languages Integration with Interactive Learning Platforms References [1] H. alharbi, “Natural language processing in arabic texts: Techniques and applications,” Journal of AI and Linguistics, 2021. [2] J. devlin, M. W. Chang, K. Lee, and K. Toutanova, “Bert: Pre-training of deep bidirectional transformers for language understanding,” in Proceedings of NAACL, 2019. [3] A. pradeep, An Introduction to NLTK and its Applications. Natural Language Processing, 2020. [4] Z. zhao, Y. Xie, and W. Tang, “Sentence segmentation and tokenization for arabic texts,” Computational Linguistics Journal, 2021. [5] A. alotaibi, “Arabic linguistics and text analysis: Challenges and solutions,” Linguistics Studies, 2022. [6] S. yoon, “gtts: Google text-to-speech for various applications,” Speech Processing, 2020. [7] R. sharma and A. Agarwal, “Automated question generation using ai: A survey,” International Journal of AI Research, 2020. [8] A. rosenfeld and P. Kaskie, “Chatbots in education: Using ai to foster interaction,” AI in Education Journal, 2020. [9] H. li and Y. Zhang, “Intelligent question generation for e-learning: Techniques and applications,” E-learning Technologies, 2020. [10] V. solanki and S. Patel, “Ai-driven educational tools for interactive learning,” AI and Learning Journal, 2021. [11] M. abdul, “Ensuring text logic and syntax in arabic sentences,” Arabic Linguistics Review, 2020. [12] M. rahman, “Word classification in arabic language using nlp,” Arabic Computational Linguistics, 2020. [13] A. Name, “Title of the paper,” Journal Name, vol. Volume Number, p. Page Range, 2023. [14] Smith, J., & Brown, A., “Title of the paper,” Journal of Arabic Studies, vol. 15, pp. 123-135, 2024. Thank You image1.png image2.svg image3.gif image4.png image5.svg image6.png image7.svg image8.png image9.svg image10.png image11.svg image12.png image13.svg image14.png image15.png image16.svg image17.png image18.svg image19.png image20.svg image21.png image22.svg image23.png image24.svg image25.png image26.svg image27.png image28.svg image29.jpeg image30.png image31.svg image32.png image33.svg /docProps/thumbnail.jpeg