Advanced Analytics and Deep Learning Models. Группа авторов

Чтение книги онлайн.

СКАЧАТЬ

In this context of language learning and teaching, it is worth noting that Warschauer and Kern coined the concept of Network-Based Language Teaching (NBLT), which focuses on communication [13]. According to Sharples et al., there is a fundamental correlation between learner-centered, personalized, interactive, collaborative, situated, lifelong, and ubiquitous New Learning and New Technology, which is well known for being mobile, user-friendly, and ubiquitous [14]. According to Jones, Networked Learning (NL) has emerged as a significant paradigm in which ICT is used to foster interaction and connections between teachers and learners, learners and other learners, and a learning community and its learning resources [15]. Further, with the Fourth Industrial Revolution which features rapid expansion of technology and digital application is influencing all spheres of our lives. AI has made an impact on the way we live and work, that is, from floor cleaning, using automatic induction heaters, driverless cars to instructing Alexa. According to Manns, the Fourth Industrial Revolution is being driven by the integration and amplification of emerging breakthroughs in AI, automation, and robotics, as well as the far-reaching connection between billions of people with mobile devices that provide unparalleled access to data and information [16]. Furthermore, AI now has major applications for language studies, thanks to advances in NLP, the advent of NL, and the technological ability to manage large amount of data.

1.3 Defining Artificial Intelligence

      Artificial Intelligence (AI) is a branch of science that studies and develops devices aimed at stimulating human intelligence processes. The primary aim of AI is to improve the speed and efficacy of regular processes. As a result, the number of industries implementing AI is growing globally [17].

      The term AI as defined by Russell and Norvig is Computational Intelligence, or Machine Intelligence, which encompasses a wide range of subfields in which “specific tasks, such as playing chess, proving mathematical theorems, writing poetry, and diagnosing diseases, can be performed” [18]. According to Housman, “AI is capable of two things: (1) automating repetitive tasks by predicting outcomes on data that has been labeled by human beings, and (2) enhancing human decision-making by feeding problems to algorithms developed by humans” [19]. To put it another way, AI registers assigned commands by performing the tasks repeatedly and then generates a decision pathway for humans by presenting alternatives. Moreover, Nabiyev describes AI as a computer-controlled device’s ability to execute tasks in a human-like manner [20]. According to the author, human-like features include mental processes like reasoning, meaning formation, generalization, and learning from prior experiences. Nilsson goes on to describe AI as the full algorithmic edifice that mimics human intellect [21]. According to him, AI encompasses the development of the information-processing theory of intelligence.

1.4 Historical Overview of AI in Education and Language Learning

AI has evolved in terms of its philosophical approach over time. Intelligent Tutoring Systems (ITSs) were the first to incorporate AI into language learning in the 1980s aimed for personalized and autonomous learning. Early iterations of ITS were referred to as programs that sought to cater to the needs of learners by facilitating communication [22]. Another significant benefit of ITS was that it allowed for infinite repetitions and practice, something that could never be done with a human instructor. It was designed for the individual learner who wanted to improve their language skills by using tutoring systems. Despite of its advantages, several studies on integration of ITS in higher education found that it had moderate positive impact on the academic learning of college students [23]. However, after four decades, the more advanced and updated version of AI has revitalized the potential for personalized learning [24].

      Although ITS made extensive use of drill and rote-learning mechanism built into the computer-based learning system, today’s AI applications are much more advanced, with the same aim of catering to personalized learning. The fundamental difference between the previous model of ITS and the current model is that the former involved a student working in isolation using an ITS and the later engages students in a networked environment. This exposes the learner to the authentic and natural learning scenarios providing social context for language learning.

      As mentioned earlier, the remarkable advancement in AI has brought a significant and inevitable shift from CALL to ICALL. With advancements in mobile technologies and their applications in language learning, CALL paved the way for MALL, and similarly, development in AI has led to the rise of a new academic field called ICALL. NLP technologies’ language processing capabilities have numerous implications in the field of CALL, and the field of study that investigates and integrates such implementations is known as ICALL [25].

      In the early 2000s, the Massive Open Online Courses (MOOCs) offered a highly required and cost effective alternative to the expensive higher education in the US and beyond. However, such courses could not facilitate learners’ participation, peer learning, scaffolding, or large-scale connections with global learners. Because of these constraints, the MOOC movement has stalled when it comes to delivering education on a wide scale. In contrast, many well-known ongoing MOOC initiatives, such as Coursera, Khan Academy, Udemi, EdX, and Udacity, have used AI and NLP techniques to improve learners’ engagement, active learning, and autonomy. This resurgence of AI, along with its strong NLP potential, has had a significant impact on second language education, as NLP-based tutoring systems can provide corrective input and adapt and customise instructional materials [5].

1.5 Implication of Artificial Intelligence in Education

There are myriad of implications of AI in language teaching and learning. There is multitude of ways that language learners and teachers can gain from integrating this technology. Some of the most relevant implications are the following.

      1.5.1 Machine Translation

      Cultural variation is one of the predominant barriers of communication which majorly occurs due to the difficulty in decoding the language, one is not familiar with. In such scenario, being bilingual or multilingual is a blessing which paves the way for enormous career opportunities and communication across the world. The language barrier is easily eradicated by innovative AI-based translation technologies like Google Translate. On a wide scale, such innovations have made significant progress in helping second language and foreign language learners. Google Translate initially supported only a few languages, but by 2016, it supported 103 languages at different levels, with over 500 million total users and over 100 billion words translated daily [26]. Since this translation service is so easily and widely accessible, second language learners are using it to enhance their learning beyond the four walls of the classroom. In contrast, Google’s machine translation had been slammed for its accuracy because the translations are based on statistical machine translation rather than grammatical rules. Advanced and revised versions of Google Translate, on the other hand, exhibited higher accuracy [27].

      1.5.2 Chatbots

      Learners can communicate and learn from language chatbots in a natural way by integrating chatbots in mobile apps, which enhances the autonomy of the learning process. Duolingo is the most common language learning chatbot, with AI algorithms that can understand the context of use and respond contextually and uniquely to users. Chatbots have helped thousands of learners learn languages without being embarrassed or feel uncomfortable. There are other such language learning chatbots like Andy, Mondly, and Memrise.

Figures 1.1 and 1.2 show how the chatbots respond to users contextually and uniquely.

      1.5.3 Automatic Speech Recognition Tools

      The speech recognition tools identify spoken languages, analyze СКАЧАТЬ