July 13 ~ 14, 2024, Virtual Conference
Jili Pu, Digital Business School，Yunnan Economics Trade And Foreign Affairs College Yunnan Kunming, 650000 China
In the wake of rapid advancements in artificial intelligence (AI), the field of education stands at the cusp of a transformative era marked by unprecedented opportunities and challenges. Information technology (IT) courses, integral to modern pedagogy, play a pivotal role in nurturing students' IT competencies and fostering a spirit of innovation. However, traditional classroom teaching methods often grapple with inefficiencies and the difficulty of catering to individual educational needs. The rise of online learning presents a novel avenue for IT course instruction, and integrating AI with these courses paves the way for an innovative teaching paradigm. This paper endeavors to architect an AI-based online teaching system for IT courses that promises an efficient, personalized, and interactive learning experience. By harnessing AI technologies and aligning them with IT curricula, the system is designed to facilitate intelligent teaching, adaptive learning, and personalized tutoring. The research delves into the untapped potential of AI in IT education and scrutinizes critical aspects of system design, encompassing teaching content management, learning path planning, knowledge recommendation, and student assessment. Through empirical studies and practical applications, the feasibility and efficacy of the proposed system are validated. The findings of this study are expected to offer an innovative approach and toolkit for IT course instruction, providing students with a more efficient, personalized, and interactive learning experience. Furthermore, by leveraging the strengths of AI, the design of smarter and more personalized teaching models can offer valuable insights and experience for educational reform and development.
Artificial Intelligence, Information Technology, Online Teaching, Teaching Systems, Personalized Learning.
Hani Almaqtari , Zeng Feng, Ammar Mohammed, School of Computer Science and Engineering, Central South University,Changsha, Hunan, 410083, China
Nowadays, many people use the internet and social media to communicate and express opinions on various topics across different aspects of life. Sentiment analysis analyse people's opinions and serves as a powerful decision-making tool for individuals and companies. Despite the growing importance of understanding sentiment in Arabic text still presents challenges due to the language's intricate nature, encompassing diverse dialects, writing styles, and linguistic nuances. To address this gap, this study proposes a novel Multi-channel Convolutional Neural Network - Bidirectional Gated Recurrent Unit (AMCB) model with Bidirectional Encoder Representations from Transformers (AraBERT), tailored to effectively capture both local and global dependencies within Arabic textual data. This approach utilizes Arabic BERT pre-trained models and tokenizers structured into three stages. The first stage involves text preprocessing and data cleaning. Following this, a pre-trained BERT model is employed to obtain sentence-level semantics and contextual features, generating embeddings. These contextual embedded vectors are then passed to the neural network. Moreover, sentiment analysis is enhanced by incorporating diverse word embedding techniques such as Word2Vec and FastText, along with advanced architectures including multi-convolutional Neural Networks and Bidirectional Gated Recurrent Unit, framework. Each of these architectures is meticulously crafted to enhance model generalization and accurately process sequential and contextual text data crucial for sentiment analysis in Arabic. Moreover, the AMCB model is evaluated across two different datasets containing reviews in various Arabic dialects and compared to state-ofthe-art models based on accuracy, precision, recall, F1-score, and AUC values. Through extensive experimentation and model optimization, the effectiveness of the AMCB model is demonstrated. The model achieved accuracies of 96.21%, and 90.94% on the (Hotel Arabic Reviews Dataset) HARD and (The Large-Scale Arabic Book Reviews) LABR respectively.
Arabic Sentiment analysis, Deep Learning , Nuatrual Lunguage Processing ,BiGRU, AraBERT.