Artificial Intelligence is drawing a lot of interest from corporates, start-ups and the government because of its potential for business and social transformation. Indian AI sector has seen 18% YoY growth in 2018 and NITI Aayog has also identified five prioritysectors— health, education, agriculture, smart cities and smart mobility in its research paper on National AI Strategy.
Globally, the total economic value derived from AI is predicted to reach USD 3.9 Trillion by 2022. Hence, this is the right time for innovation in AI-based applications for improving the quality of life of rural population of India, especially in primary sectors. TEACHER SHORTAGE Indian village schools suffer from persistent shortage of qualified teachers. Globally, there is a need for 24.4 million primary school teachers to achieve UNESCO’s goal of universal primary education by 2030. In India, although the overall pupilteacher ratio (PTR) has been improving gradually, most of the teachers in rural areas lack proper training.
This is further aggravated by the fact that India has nearly 100,000 schools, mostly in villages, which are single-teacher schools. This impacts PTR negatively and leads to deterioration of learning experience of the students eventually affecting their motivation to attend classes regularly. Under these circumstances, teaching of all subjects, including English, has suffered a lot across schools in villages. This vacuum in rudimentary education is very difficult to fill using skill-based trainings done later. Unless this conundrum is resolved, ensuring employability of a huge segment of population, is going to be a persistent challenge. Hence, there is an urgent need to augment the teaching community at village schools using digital technologies as well as AI . LANGUAGE TEACHING English has emerged as universally accepted “lingua-franca” globally as well as across India. But various surveys have revealed the state of English language skills of Indian students. Especially, among rural students, the situation is worse, and it affects their employability in various customer facing roles across industries. Teaching of English can benefit from various capabilities of NLP (Neuro Linguistic Programming) technique of AI to make it more effective. On one hand, it helps the students to pick up right skills at right pace, and on the other hand, it exposes teachers to emerging digital technologies. WRITING SKILLS Various NLP-based techniques, generated using (machine learning) ML and word processing, can be used to automate the process of teaching writing English to students. For example, automatic detection of incorrect spellings/punctuations, grammatically incorrect phrases or sentences, generating reward points for correctly formed sentences and such. The process can be made more engaging by starting with partially filled paragraph templates on chosen topics, where students work on correctly guessing and filling the blanks by adding words, phrases, sentences appropriately. Another advanced technique namely, words/phrase prediction on partially constructed sentences can help students to fill the gaps/void which they face while writing a paragraph on some topic. Such a model of Automated Assisted Learning (AAL) enables more experimentation, innovative thinking. Even for translation, the process that is a part of a student understanding the subjects, there are NLP models trained on language translation, which includes the relevant dictionary mapping (e.g. Hindi->English). SPEAKING AND READING For most Indian children, spoken English practice follows learning of writing skills. It is also true that many students who acquire good writing skill, still fear speaking in English, which is perceived as essential skill by employers. One way to help young people to come out of this is to engage them in regular English conversation. A chatbot, like Google Personal Assistant, trained on topic-based conversation can really help students to learn speaking English. To start with, the student can try pronouncing individual words correctly (by listening to pronunciation of the chatbot), followed by speaking individual phrases and complete sentences.
As availability of appropriate corpus is an important pre-requisite of building such a learning application, initially such a model can be targeted for teaching English to students speaking most common languages, such as Hindi, Marathi, Bengali, Punjabi or Tamil.
Source: Hindustan Times dated 27 March, 2019