Over the last one year, thanks to massive disruption of physical teaching, Edtech innovators have shifted their focus to building a deployable plug and play system- which can adapt to any environment, such as
1. Low Infrastructure Requirement: Low teaching hardware requirements simplify the on-ground operation, with no need for an internet connection i.e., it works in offline mode as well. Learning can start with only one tab.
2. Fitment across all platforms and Operating systems: By being platform and device agnostic, teachers and students will be able to learn at any given time. Innovations include the development of a special AI-based Teaching System that has different versions for Windows, iOS, and Android. The PC version is set up at the kindergarten teaching centres and/or schools, while parents and teachers and parents can download the mobile version directly. That way, students can learn both at school and at home using the tool.
3. Adaptive learning: Thanks to the application of AI, two children in the same classroom can learn at a different pace, as per their grasp and comfort. This also applies to the child’s pre-assessment, continuous teacher training via chatbots, additional provision of offline learning tools and materials and assessment of learning objectives.
4. Customisable Solutions: Organizations must understand the diverse needs of a country like India, where each state boasts a different learning requirement. The need of the hour is to draw from the best of globally-aligned practices and create programs that are tailored to the Indian context. Only by customising solutions to suit learning needs in each state will speedy and smooth on-ground implementation be assured.
As schools, universities, and institutions are pondering “what next” amid global lockdowns, the world opened up to the many intriguing possibilities and potential of digital learning. AI has the potential to give children the opportunity to build their literacy, numeracy, digital, and life skills to succeed.
AI can be reimagined for education in the following ways-
Knowledge space theory: Gaps in students’ knowledge can be identified by using Knowledge Space Theory to train AI learning platforms. This theory uses mathematical language to pinpoint and track “knowledge points”, that when put together offer a precise picture of an individual’s “knowledge state” for any given subject. This helps to improve the efficiency and personalization of learning, such as helping to develop curricula tailored to each student’s specific needs.
Smart content: AI can allow for real-time personalization of content that ensures each lesson is delivered to learners in a manner that is best suited to them, be it through gamification and simulations, virtual courses, self assessments, or other tools. One way of doing so is algorithmic adaptivity to understand what the learner knows, and deliver precise curriculum tailored to a learner’s current knowledge base. An algorithmic approach enabling this is Bayesian Knowledge Tracing, which helps estimate rates of knowledge or learning absorption. Another framework of value is Item Response Theory (IRT), often used in psychometrics to model a learner’s interactions with discrete items. These can help generate content updates dynamically for multiple learning curves, while also helping to greatly improve content discovery and engagement.
Predictive testing and assessment: The use of AI in assessment is not new: expert algorithms have been used for years in psychometric assessments. Now, we’re seeing AI being used to generate unique test questions in real-time, and being applied in adaptive scoring. One such example of this is Computerized Adaptive Testing (CAT), which implements predictive testing adjusted to an examinee’s skill level via dynamically selected appropriate testing items. These work through the creation of comprehensive item banks which contain all the items that could appear in the test, which must first be calibrated through pre-testing.
Using this data, IRT is then used to calculate parameters for each of the items in the bank, which includes the difficulty of the item, including, minimum and maximum number of items to present, stopping conditions, or more. Where these differ from traditional testing is that an algorithm selects the next item in real time, ensuring controlled coverage of all content areas across the bank as a whole. This is a significant move away from the one size fits all school of education that is so commonplace, and allows educators to instead tailor testing and assessment mechanisms to the aptitude of students that are at different levels of a learning curve. This increases the chances of learning retention and absorption of knowledge and skills by children, as the system adapts to their unique needs.
Natural Language Processing (NLP): Curiosity and inquisitiveness can be nurtured with chatbots based on Natural Language Processing (NLP) to solve the queries of young learners, parents, and teachers. By using machine learning to parse inputs and interact, we can use technology to interact in a more sympathetic, human manner, thus improving learning experiences while keeping people actively engaged. Using computational algorithms, context extraction, content summary, and sentiment examination, NLP can help chatbots decipher context, intent, and sentiment and thereby deliver enhanced educational outcomes.
What is helping is that the Indian government is investing more than Rs 1 Lakh Cr. in RISE (Revitalising Infrastructure and Systems in Education), which focuses on embedding new-age technologies such as cloud computing, AI and VR into our education system. By making the most of the cloud, we can continue onboarding, teaching, and engaging a large slice of the student population no matter where they might be. Students and teachers can make the most of robust, localised, and contextualised Edtech platforms to enable interactive learning and bridge the physical divide. By creating a strong tech back-end, a thriving app ecosystem, and strategic API integrations, young learners can access remotely at any given time, all through the magic of the internet.
The author is Managing Director, Square Panda India