Publications - AmritaCREATE

There is an explosive growth in e-Learning platforms, jointly developed by multiple institutions, which provide for virtual learning content. However, many are inadequate to support the complex requirements for collaborative development of distributed learning such as accommodation of wide-ranging technologies, servers, and remote equipment controlled by diverse software. Our solution is a multi-tier architecture that supports collaborative development, publishing in various online and print formats, security, audit, and access controls. Our design considerations include a highly scalable platform, use of open technologies, templates that provide pedagogical structure, multilingual functionality, and shared virtual availability of lab equipment from multiple geographic locations, along with secure access to remote equipment.

The Cognitive Theory of Multimedia Learning (CTML) offers insights on creating, organizing and presenting multimedia content in a way that can enhance the effectiveness of learners. While several studies have been conducted to verify these principles, not many have been done using content presented on low cost tablets. In this paper, we present our experimental study on the retention and understanding by children when learning content a) adheres to CTML and b) offered on low cost tablets. In our study we used Online Labs theory and videos and incorporated CTML principles into it. According to CTML principle it will maximize learning without causing cognitive overload. The study sample was 58 students from an English medium secondary education school from south India. The experiment involves a control and experiment group where the control group is tested on content which does not adhere to CTML and the experimental group is tested on content designed based on CTML. A pre-test is conducted to assess the level of pre-knowledge in the subject and a post-test is conducted to assess the retention and understanding by students. This suggests that tablet content incorporated with CTML principle can lead to a deeper understanding of the subject.

Purpose: The purpose of this paper is to illustrate an Inclusive Digital Literacy Framework for vulnerable populations in rural areas under the Digital India program. Key challenges include addressing multiple literacies such as health literacy, financial literacy and eSafety for low-literate learners in low-resource settings with low internet bandwidth, lack of ICT facilities and intermittent electricity. Design/methodology/approach: This research implemented an educational model based on the proposed framework to train over 1,000 indigenous people using an integrated curriculum for digital literacies at remote settlements. The model uses mobile technology adapted for remote areas, context enabled curriculum, along with flexible learning schedules. Findings: The education model exemplifies a viable strategy to overcome persistent challenges by taking tablet-based digital literacies directly to communities. It engages different actors such as existing civil societies, schools and government organizations to provide digital literacy and awareness thereby improving both digital and life skills. It demonstrates the potential value of a comprehensive Digital Literacy framework as a powerful lever for Digital Inclusion. Practical Implications: Policy makers can use this transformational model to extend the reach and effectiveness of Digital Inclusion through the last mile enhancing existing training and service centers that offer the traditional model of Digital Literacy Education. Originality/value: This innovative mobile learning model based on the proposed Digital Framework for Inclusion instilled motivation, interest and confidence while providing effective digital training and conducting exams directly in the tribal settlements for low-literate learners in remote settings. Through incorporating multiple literacies, this model serves to empower learners, enhance potential, improve well-being and reduce the risk of exploitation. © 2018, Emerald Publishing Limited.

Purpose of Review: This article discusses the advances, methods, challenges, and future directions of data-driven methods in advancing precision oncology for biomedical research, drug discovery, clinical research, and practice. Recent Findings: Precision oncology provides individually tailored cancer treatment by considering an individual’s genetic makeup, clinical, environmental, social, and lifestyle information. Challenges include voluminous, heterogeneous, and disparate data generated by different technologies with multiple modalities such as Omics, electronic health records, clinical registries and repositories, medical imaging, demographics, wearables, and sensors. Statistical and machine learning methods have been continuously adapting to the ever-increasing size and complexity of data. Precision Oncology supportive analytics have improved turnaround time in biomarker discovery and time-to-application of new and repurposed drugs. Precision oncology additionally seeks to identify target patient populations based on genomic alterations that are sensitive or resistant to conventional or experimental treatments. Predictive models have been developed for cancer progression and survivorship, drug sensitivity and resistance, and identification of the most suitable combination treatments for individual patient scenarios. In the future, clinical decision support systems need to be revamped to better incorporate knowledge from precision oncology, thus enabling clinical practitioners to provide precision cancer care. Summary: Open Omics datasets, machine learning algorithms, and predictive models have enabled the advancement of precision oncology. Clinical decision support systems with integrated electronic health record and Omics data are needed to provide data-driven recommendations to assist clinicians in disease prevention, early identification, and individualized treatment. Additionally, as cancer is a constantly evolving disorder, clinical decision systems will need to be continually updated based on more recent knowledge and datasets.

Educational schemes for marginalized communities such as SC/ST were introduced by the Government of India with the aim of providing equal opportunity for scheduled tribe and scheduled caste to empower them with free educational facilities. For such programs to be effective, SC/ST members need to be aware of them and need to have a positive attitude towards them. However, due to the practice of untouchability and the remoteness of their dwellings, SCs and STs have suffered isolation from mainstream population for several centuries. As a consequence, they continue to suffer from a high degree of educational exclusion. Upliftment of SCs and STs, both economically and socially, will be only possible through inclusive education efforts. This study focus on the role ofawareness on educational schemes in attaining primary and secondary education among SC/ST communities in two rural villages in Coimbatore. The major findings of the study reveals that most of marginalized communities continue to remain unaware of the educational schemes and are unable to utilize all the provisions of the schemes. © 2018 Academic Press.

Rural India lacks easy access to health practitioners and medical centers, depending instead on community health workers. In these areas, common ailments that are easy to manage with medicines, often lead to medical escalations and even fatalities due to lack of awareness and delayed diagnosis. The introduction of wearable health devices has made it easier to monitor health conditions and to connect doctors and patients in urban areas. However, existing initiatives have not succeeded in providing adequate health monitoring to rural and low-literate patients, as current methods are expensive, require consistent connectivity and expect literate users. Our design considerations address these concerns by providing low-cost medical devices connected to a low-cost health platform, along with personalized guidance based on patient physiological parameters in local languages, and alerts to medical practitioners in case of emergencies. This patient-centric integrated healthcare system is designed to manage the overall health of villagers with real-time health monitoring of patients, to offer guidance on preventive care, and to increase health awareness and self-monitoring at an affordable price. This personalized health monitoring system addresses the health-related needs in remote and rural areas by (1) empowering health workers in monitoring of basic health conditions for rural patients in order to prevent escalations, (2) personalized feedback regarding nutrition, exercise, diet, preventive Ayurveda care and yoga postures based on vital parameters and (3) reporting of patient data to the patient's health center with emergency alerts to doctor and patient. The system supports community health workers in the diagnostic procedure, management, and reporting of rural patients, and functions well even with only intermittent access to Internet.

In remote rural areas, the availability of trained and qualified teachers is poor. Rural teaching is teachercentric instead of student-centric. A systematic monitoring system to ensure that teachers practice the methods they learn at pre- or in-service trainings is not yet in place. Plagued by such deficiencies, present day rural education in India lacks quality, and children read, write, and compute several levels below grade level. Our intervention comes from the premise that reform approach in rural education should be at the grassroots level, involving comprehensive change in teaching and learning attitudes, teaching methodology and subject matter expertise. Amrita Rural India Tablet Education Teacher Training Program is a technology-enabled training program for teachers that train them to be change-agents and influence both children and communities to bring about sustainable change in attitudes as well as education and learning outcomes. AmritaRITE utilizes cutting edge mobile learning technology in both teaching and monitoring, to improve quality and accountability, even in low-electricity and intermittent-connectivity areas. The comprehensive program trains teachers in tablet-supported teaching methodologies, multi-age classrooms, classroom management styles suited to specific environment, addressing different learning modalities, developing student critical thinking, identifying learning disabilities and overcoming social barriers such as educating girls and low literacy communities.

Medical Virtual Patient Simulator (MedSIM) is a cutting-edge E-learning innovation for medical and other health professionals. It consists of a framework that supports various patient cases, tailored by interdisciplinary medical teams. Each virtual patient case follows the critical path to be followed for a specific patient in a hospital. MedSIM takes the student on a journey that enables the student to interview, examine, conduct physical, systematic and ultimately reach a diagnosis based on the path that is chosen.

The problem of high dimensionality in multi-label domain is an emerging research area to explore. A strategy is proposed to combine both multiple regression and hybrid k-Nearest Neighbor algorithm in an efficient way for high-dimensional multi-label classification. The hybrid kNN performs the dimensionality reduction in the feature space of multi-labeled data in order to reduce the search space as well as the feature space for kNN, and multiple regression is used to extract label-dependent information from the label space. Our multi-label classifier incorporates label dependency in the label space and feature similarity in the reduced feature space for prediction. It has various applications in different domains such as in information retrieval, query categorization, medical diagnosis, and marketing.

Matrix factorization is the most popular approach to solving prediction problems. However, in the recent years multiple relationships amongst the entities have been exploited in order to improvise the state-of-the-art systems leading to a multi relational matrix factorization (MRMF) model. MRMF deals with factorization of multiple relationships existing between the main entities of the target relation and their metadata. A further improvement to MRMF is the Weighted Multi Relational Matrix Factorization (WMRMF) which treats the main relation for the prediction with more importance than the other relations. In this paper, we propose to enhance the prediction accuracy of the existing models by personalizing it based on student knowledge and task difficulty. We enhance the WMRMF model by incorporating the student and task bias for prediction in multi-relational models. Empirically we have shown using over five hundred thousand records from Knowledge Discovery dataset provided by Data Mining and Knowledge Discovery competition that the proposed approach attains a much higher accuracy and lower error(Root Mean Square Error and Mean Absolute Error) compared to the existing models.

The problem of high dimensionality in multi-label domain is an emerging research area to explore. A strategy is proposed to combine both multiple regression and hybrid k-Nearest Neighbor algorithm in an efficient way for high-dimensional multi-label classification. The hybrid kNN performs the dimensionality reduction in the feature space of multi-labeled data in order to reduce the search space as well as the feature space for kNN, and multiple regression is used to extract label-dependent information from the label space. Our multi-label classifier incorporates label dependency in the label space and feature similarity in the reduced feature space for prediction. It has various applications in different domains such as in information retrieval, query categorization, medical diagnosis, and marketing.

Medical Virtual Patient Simulator (MedVPS) is a cutting-edge eLearning innovation for medical and other health professionals. It consists of a framework that supports various patient cases, tailored by interdisciplinary medical teams. Each virtual patient case follows the critical path to be followed for a specific patient in a hospital. MedVPS takes the student on a journey that enables the student to interview, examine, conduct physical, systematic and ultimately reach a diagnosis based on the path that is chosen. After the interactions, the student must decide whether each response is normal or abnormal and use the virtual findings to identify multiple probable diagnoses or reexamine the virtual patient with the goal to narrow down to the correct disease and then provide treatment. We present the architecture and functionality of the MedVPS platform and include a pilot study with medical students. © Springer International Publishing Switzerland 2016.

ICT-enabled virtual and remote labs have become a platform augmenting user engagement in blended education scenarios enhancing University education in rural India. A novel trend is the use of remote laboratories as learning and teaching tools in classrooms and elsewhere. This paper reports case studies based on our deployment of 20 web-based virtual labs with more than 170+ online experiments in Biotechnology and Biomedical engineering discipline with content for undergraduate and postgraduate education.

This paper examines the dynamics of access and exclusion in children’s Internet use, in both private and public school spaces and interrogates the role of socioeconomic and demographic predictors as well as the schooling system in shaping Internet habits. More specifically, it explores the nature of Internet use by primary school children, mainly for education and information and attempts to understand the differences across and within two types of schools- a rural public school and an elite private school. Through in-depth interviews, this research investigates the level of computer and Internet literacy among the primary school children in the age group of 8-10 years and reports the differences observed among the various social dimensions. It attempts to stress the significance and need in today’s context to provide the opportunities for physical and material access so that disadvantaged children are not excluded from the digital opportunities. © Media Watch.

Personalization and adaptation are at the core of Intelligent Tutoring Systems. The Bayesian Knowledge Tracing (BKT) Student Model is a time-tested method that maintains information about students' knowledge levels for the different skills in the topic domain. In our previous work, we had proposed the Personalized, Clustered, Bayesian Knowledge Tracing (PC-BKT) model that individualizes the learning of skills for each student and additionally improves the prediction for the cold start problem. A clustering of both students and skills based on a student and skill capability matrix was used to learn the prior skills to deal with the cold start problem, which is the prediction for either new skills or new students. Both the BKT and the PC-BKT models assume that a skill once learnt is never forgotten. But forgetting is pervasive. If a previously learnt skill is not used for a while, there is a higher chance of forgetting it. One of the factors that influence the forgetting is the time duration before the current attempt at using a skill and the previous attempt. We incorporate forgetting as a time decay function in the BKT and PC-BKT models and show significant increase in the accuracy of the student prediction.

An Intelligent Tutoring System (ITS) supplements traditional learning methods and is used for personalized learning purposes that range from exploring simple examples to understanding intricate problems. The Bayesian Knowledge Tracing (BKT) model is an established method for student modeling. A recent enhancement to the BKT model is the BKT-PPS (Prior Per Student) which introduces a prior learnt for each student. Although this method demonstrates improved prediction results compared to the others, there are several aspects that limit its usefulness; (a) for a student, the prior learning is common for all skills, however in reality, it varies for each skill (b) Different students have varying learning capabilities; therefore these students cannot be considered as a homogenous group. In this paper, we aim to improve the prediction of student performance using an enhanced BKT model called the PC-BKT (Personalized & Clustered) with individual priors for each student and skill, and dynamic clustering of students based on changing learning ability. We evaluate the predictions in terms of future performance within ASSISTments intelligent tutoring dataset using over 240,000 log data and show that our models increase the accuracy of student prediction in both the general and the cold start problem.

Our paper proposes a new pedagogical approach for learning chemistry practical experiments based on three modes of inquiry-based learning namely; structured, guided and open. Online Labs (OLabs) is a web-based learning environment for science practical experiments that include simulations, animations, tutorials and assessments. Inquiry-based learning is a pedagogy that supports student-centered learning and encourages them to think scientifically. It develops evidence based reasoning and creative problem solving skills that result in knowledge creation and higher recall. We discuss the methodology and tools that OLabs provides to enable educators to design three types of inquiry-based learning for Chemistry experiments. The integration of inquiry-based learning into OLabs is aligned with the Indian Central Board of Secondary Education (CBSE) goal of nurturing higher order inquiry skills for student centered and active learning. Inquiry-based OLabs pedagogy also empowers the teachers to provide differentiated instruction to the students while enhancing student interest and motivation.

The reliability of automated essay scoring (AES) has been the subject of debate among educators. Most systems treat essays as a bag of words and evaluate them based on LSA, LDA or other means. Many also incorporate syntactic information about essays such as the number of spelling mistakes, number of words and so on. Towards this goal, a challenging problem is to correctly understand the semantics of the essay to be evaluated so as to differentiate the intended meaning of terms used in the context of a sentence. We incorporate an unsupervised word sense disambiguation (WSD) algorithm which measures similarity between sentences as a preprocessing step to our existing AES system. We evaluate the enhanced AES model with the Kaggle AES dataset of 1400 pre-scored text answers that were manually scored by two human raters. Based on kappa scores, while both models had weighted kappa scores comparable to the human raters, the model with the WSD outperformed the model without the WSD.

Feature selection algorithms look to effectively and efficiently find an optimal subset of relevant features in the data. As the number of features and the data size increases, new methods of reducing the complexity while maintaining the goodness of the features selected are needed. We review popular feature selection algorithms such as the probabilistic search algorithm based Las Vegas Filter (LVF) and the complete search based Automatic Branch and Bound (ABB) that use the consistency measure. The hybrid Quick Branch and Bound (QBB) algorithm first runs LVF to find a smaller subset of valid features and then performs ABB with the reduced feature set. QBB is reasonably fast, robust and handles features which are interdependent, but does not work well with large data. In this paper, we propose an enhanced QBB algorithm called QBB Iterative (QBB-I).QBB-I partitions the dataset into two, and performs QBB on the first partition to find a possible feature subset. This feature subset is tested with the second partition using the consistency measure, and the inconsistent rows, if any, are added to the first partition and the process is repeated until we find the optimal feature set. Our tests with ASSISTments intelligent tutoring dataset using over 150,000 log data and other standard datasets show that QBB-I is significantly more efficient than QBB while selecting the same subset of features. © Springer International Publishing Switzerland 2014.

Virtual Labs (VLAB) is a multi-institutional Open Educational Resources (OER) initiative, exclusively focused on lab experiments for engineering education. This project envisages building a large OER repository, containing over 1650 virtual experiments mapped to the engineering curriculum. The introduction of VLAB is a paradigm shift in an educational system that is slow to change. Treating VLAB OER as an educational technology innovation, its adoption by potential-adopter engineering students (N=131) is modeled based on Roger's theory of perceived attributes. Regression and factor analysis were used to analyze the data. Results indicate that the attributes of Compatibility, Ease of Use, Relative Advantage, and Trialability significantly influence potential-adopter students' intention to adopt an innovation like VLAB. The study also observed that using OER (such as VLAB) on desktops and low-cost tablets had similar effects in student performance to using physical labs. This has interesting implications for education policy-makers who are looking to reduce the digital divide.

Within the context of Roger’s Diffusion of Innovation theory we propose a pedagogical framework for attributes that can significantly affect student adoption of collaborative learning environment like multi-user, multi-touch tabletop. We investigated the learning outcomes of secondary school students in India collaboratively using OLabs on a tabletop (EG1 = 30) vs. individually using at desktops (EG2 = 92). We analyzed the nature of communication, touch and non-touch gesture actions, position around the tabletop, focus group interviews, and pre and post test scores. Using Bass model the study also accounts for the inter influence of related group of potential adopter teachers who are likely to exert positive influence on students. The results revealed that learning outcomes on tabletop are strongly associated with innovation attributes like Relative Advantage, Compatibility, Ease of Use, Perceived Enjoyment, Perceived usefulness and Teachers support. Overall students expressed much more positive attitude to adopt tabletop technology for learning vs. desktop. We find that the mean group performance gain is significant with collaboration using tabletop and significantly greater than the group using desktops. We also find that the group interactions with the tabletop area significant factor that contributes to the group's average performance gain. However, the total time spent in while using the tabletop is surprisingly not a significant factor in the performance gain. Our findings contribute to the design of new pedagogical models for science learning that maximizes the collaborative learning potential of tabletops.

 

Keywords: Laboratory, Diffusion, Innovation, Simulation, Collaboration, tabletop, experiments

While there is large body of work examining efficacy of Virtual Labs in engineering education, studies to date have lacked modeling Blended Labs (BL) – mix of Virtual Labs (VL) and Physical Labs (PL) for science experimentation at the university engineering level. Using Rogers theory of perceived attributes, this paper provides a research framework that identifies the attributes for BL adoption in a social group comprising of (N=246) potential adopter undergraduate engineering students. Using Bass model the study also accounts for the interinfluence of related group of potential adopter faculties who are likely to exert positive influence on students. The results revealed that acceptance of BL as an innovation and its learning outcomes are strongly associated with innovation attributes like Relative Advantage, Compatibility, Ease of Use, Department and Faculty support. Learning outcomes are very positive under BL when compared to PL, though within BL, ordering of PL and VL was not significant. For certain innovation attributes gender differences were significant. Overall students expressed much more positive attitude to adopt BL model for learning than using only PL.

Keywords: Virtual Labs, Blended Learning, Innovation Diffusion, Experiments, Engineering, interinfluence

In large classrooms with limited teacher time, there is a need for automatic evaluation of text answers and real-time personalized feedback during the learning process. In this paper, we discuss Amrita Test Evaluation & Scoring Tool (A-TEST), a text evaluation and scoring tool that learns from course materials and from human-rater scored text answers and also directly from teacher input. We use latent semantic analysis (LSA) to identify the key concepts. While most AES systems use LSA to compare students’ responses with a set of ideal essays, this ignores learning the common misconceptions that students may have about a topic. A-TEST also uses LSA to learn misconceptions from the lowest scoring essays using this as a factor for scoring. ‘A-TEST’ was evaluated using two datasets of 1400 and 1800 pre-scored text answers that were manually scored by two teachers. The scoring accuracy and kappa scores between the derived ‘A-TEST’ model and the human raters were comparable to those between the human raters.

Keywords: Feature extraction, Essay scoring, text analysis, text mining, Latent Semantic Analysis (LSA), SVD, Natural Language Process- NLP, AES

India's educational challenge includes a large school going population, shortage of science teachers and lack of science labs in many schools. To counter this challenge, the Online Labs (OLabs) pedagogy is designed as a complete learning environment with tutorials, theory, procedure, animations, videos and simulations while the assessment includes conceptual, experimental, procedural and reporting skills.

We discuss two separate empirical studies using OLabs to study the performance gains, student attitudes and preferences while using physical labs, desktops and tablets. The first study was at a school that compared students who learnt individually with OLabs on desktops, to students who learnt with the traditional teacher led physical labs. The second study was at a science camp and compared OLabs on desktops to OLabs that were context adapted for android tablets. There were significant differences between the physical labs and the self study mode using OLabs on desktops, but no significant differences between OLabs on desktops compared to OLabs on tablets.

 

Keywords—Olabs; traditional labs; animation; simulation; online labs; virtual labs; elearning; cognitive learning; conceptual skills; experimental skills; procedural skills; reporting skills; assessment

Virtual labs for science experiments are a multimedia technology innovation. A possible growth pattern of the perceived critical mass for virtual labs adoption is modeled using (N=240) potential-adopter teachers based on Roger’s theory of diffusion and of perceived attributes. Results indicate that perceived critical mass influences behavior intention to adopt a technology innovation like Virtual Labs and is affected by innovation characteristics like relative advantage, ease of use and compatibility. The work presented here models the potential-adopter teacher’s perceptions and identifies the relative importance of specific factors that influence critical mass attainment for an innovation such as Virtual Labs.

 

Keywords - virtual labs; innovation diffusion; critical mass; simulation; lab experiments

A new modelling approach for diffusion of personalized learning as an educational process innovation in social group comprising adopter-teachers is proposed. An empirical analysis regarding the perception of 261 adopter-teachers from 18 schools in India about a particular personalized learning framework has been made. Based on this analysis, teacher training (TT) has been identified as one of the dominant factor which can significantly influence decision by teachers to adopt the educational innovation. Different situations corresponding to fixed and time dependent dynamic carrying capacity of potential adopter-teachers at any time have been developed. New generalized models capturing the growth dynamics of the innovation diffusion process in conjunction with the evolutionary carrying capacity of potential adopters are investigated. The coupled dynamics allows forecasting the likelihood of success or failure of new educational innovation in a given context. Different scenarios for TT are considered based on constant growth rate model; proportional growth rate model; stratified growth rate model. The proposed modelling framework would be of great interest to education policy makers as it has the potential to predict the likelihood of success or failure of new educational innovation.

 

Keywords - Teacher training ; Personalized learning; Carrying capacity ; Innovation diffusion; Dynamic model; Educational innovation

Most personalized learning systems are designed for either personal computers (e-learning) or mobile devices (m-learning). Our research has resulted in a cloud-based adaptive learning system that incorporates mobile devices into a classroom setting. This system is fully integrated into the formative assessment process and, most importantly, coexists with the present e-learning environment. Unlike many mobile learning systems, this system provides teachers with real-time feedback about individual and group learners. Its scalable and extendable architectural framework includes the server-side pedagogical recommendation of content adaptation based on the users’ knowledge-levels and preferences. Content is also automatically adapted to the end device that is being used. This context-aware delivery allows users to switch between e-learning and m-learning, and between devices, without any loss in personalized content. Our work builds on a web-based Adaptive Learning and Assessment System (ALAS) that is built on the Knowledge Space Theory model. At present, this system is used at school computer labs and our goal was to widen this user-base by enhancing this system to support personalized learning on mobile devices. This study describes our process of developing this technology, and contains an empirical analysis of students’ performance, perceptions, and achievements when using ALAS on both personal computers and mobile devices.

 

Keywords - Adaptive assessment ; Mobile learning ; Personalized learning ; Adaptive learning ; Mobile device ; Mobility ; Intelligent tutoring ; m-Learning; e-Learning