Personalized Learning with Mindspark

Despite strides in access to education (with 98% of students having a nearby primary school), a critical issue persists in India: unequal learning outcomes. The percentage of students not learning at grade level in India and globally is a critical concern. According to the Condition of Education 2020, 20% of fourth-grade and 30% of eighth-grade students failed to reach a basic level in math on the National Assessment of Educational Progress (NAEP) in 2017. Students from socio-economically disadvantaged backgrounds, particularly those in rural and remote areas, face a multitude of challenges hindering their education.

High teacher absenteeism, a one-size-fits-all approach, rote learning, and limited appropriate support for struggling students create a vast disparity in learning levels within classrooms and across schools. This is especially true for students from marginalised families, who often lack basic foundational skills. The situation is even more stark in tribal communities, where literacy rates can be 30-40% below the national average. The pandemic further exacerbated these inequalities due to constrained access to technology, electricity and other infrastructure along with a lack tech-savviness among these disadvantaged communities.

The Root of the Problem: Traditional Methods Fall Short

Traditional education often struggles to cater to individual student needs. A single teacher managing a large class with diverse learning styles can't effectively personalize instruction. Rote learning - focused on memorization over understanding - fails to equip students with critical thinking skills. This leaves many students, especially those starting at a lower level, falling behind without the support they need to catch up.

Our Solution: Personalized Learning at Scale

Mindspark addresses these issues head-on by providing personalized adaptive learning (PAL) specifically designed for students in government schools. This innovative tool tailors learning to each student's individual needs, ensuring everyone is challenged appropriately.

Mindspark tackles the critical issue of unequal learning outcomes by providing a scalable, personalized learning solution for students from disadvantaged backgrounds. By empowering students, equipping teachers, and fostering transparency, Mindspark paves the way for a more equitable and effective education system for all.

Ei Mindspark is a personalized adaptive learning (PAL) tool that enables Teaching at the Right Level (TaRL). It is an inquiry-based and gamified software that itemizes the local syllabus. It diagnoses student’s learning levels and offers interactive questions to help students learn concepts at their own levels. These questions are based on two decades of research on common student misconceptions in languages and math. Mindspark caters for students with varying learning capacities. For students struggling with reading skills, it offers voiceovers for support, while it challenges higher-performing students with more complex questions. Mindspark also integrates AI to assist students in improving their reading comprehension and creative writing skills.

How does Mindspark work?

Empowering Students Through Individualized Learning Paths:

• Accurate Assessment: Mindspark diagnoses a student's learning level, creating a personalized starting point.

• Adaptive Content: Interactive questions, based on decades of research on common misconceptions, address specific learning gaps.

• Differentiation for All: Struggling readers receive voiceover support, while advanced learners face progressively complex challenges.

• AI-Powered Support: Mindspark leverages AI to assist students in refining reading comprehension and creative writing skills.

Equipping Teachers to Make a Difference:

• Real-Time Data Insights: Dashboards provide teachers with a clear picture of student usage, performance data, and areas requiring attention.

• Targeted Instruction: This data empowers teachers to tailor their classroom instruction and offer targeted support to struggling students.

• Focus on Learning, Not Busywork: Mindspark frees up valuable teacher time by handling individualized instruction, allowing them to focus on fostering deeper understanding.

Beyond the Classroom: A Ripple Effect

Mindspark's impact extends beyond students and teachers, fostering positive change within the education system:

• Transparency and Accountability: Usage and learning data provide school administrators with insights to identify areas needing targeted support, promoting a more transparent system.

• Data-Driven Policymaking: Valuable research data from Mindspark usage informs future educational policies and interventions.

• Prioritizing Cost-Effective Solutions: In a country like India, the need for cost-effective solutions in education has never been more critical. With the devastating learning losses caused by COVID-19, there is an urgent need to test and create models that can mitigate these effects. Mindspark stands out as a cost-effective option that not only addresses the current challenges but also provides a scalable and sustainable approach to education.

Mindspark is implemented in schools through shared computer labs, allowing students to access the program individually as part of their weekly schedules. Real-time dashboards enable teachers to monitor student usage and performance, aiding in identifying areas for improvement, which is particularly beneficial in resource-constrained public schools with large class sizes. J-PAL's RCTs have demonstrated a clear correlation between Mindspark usage and improved learning outcomes. Government stakeholders, funders also benefit from these dashboards, overseeing regional learning variations and strategizing system-wide improvements. 

Mindspark's impact extends beyond students, enhancing teacher capacity and boosting student motivation and attendance rates. It has shown effects comparable to TaRL interventions but with faster results and scalability, independent of high-quality human resources.

Mindspark is an adaptive learning solution designed to specifically address the needs of students from socio-economically disadvantaged backgrounds studying in government schools in India. These students often face a multitude of challenges that hinder their educational experience. Despite the progress made in school accessibility, significant challenges persist. High teacher absenteeism, the prevalence of rote learning, and a lack of personalized attention create a wide gap in learning levels within and between schools. This gap is most acute for students from impoverished families and those in rural or remote areas. In tribal communities, the literacy rate can be as much as 30-40% below the national average.

Mindspark tackles these challenges head-on by:

• Personalized Learning: It tailors lessons to individual student needs. This is crucial for students who may be starting from a lower baseline or learning at a slower pace.

• Teaching at the Right Level (TaRL): By adapting content to the student's current understanding, Mindspark ensures they are neither overwhelmed nor bored, fostering a more engaging learning experience.

• Actionable Insights for Teachers: Mindspark equips teachers with valuable data on student progress. This allows them to identify areas needing improvement and adapt their teaching strategies accordingly.

• Continuous Improvement: Insights from student usage and assessments are used to refine Mindspark's content and questions, while research in pedagogy further informs its development.

Mindspark's impact extends beyond students, positively affecting the education system as a whole:

• Transparency: Data from Mindspark usage helps identify schools or regions needing targeted support, leading to a more transparent and accountable education system.

• Research and Policy Development: Data generated by Mindspark is valuable for research purposes, informing future educational policies and interventions.

Mindspark revolutionizes education by breaking away from conventional methods. It employs contextualized content, using relatable examples from the learning environment of the child, to ensure concepts are easily understood. The program also engages Local Instructor Cadres (LICs) from the same community, integrating educational technology seamlessly into classrooms. Delivering content in the language of the child further enhances comprehension and engagement. Our commitment to reaching even the most remote regions is evident through its innovative measures, such as using tablets, Raspberry Pi, and portable cabins. The program currently serves 80% of its users from the most marginalized communities, with a strong emphasis on empowering girls, who make up over 60% of its user base.

Mindspark isn't just a product; it's a complete solution built by a team that comprehends the very issues it aims to address. Our 350+ members hail from diverse backgrounds in education, assessment, and large-scale government projects across India. This ensures we have a deep understanding of the education system's complexities and the challenges faced by government schools.

Our team includes teachers with years of classroom experience, curriculum and textbook developers. They understand the specific needs of students and the realities of resource-constrained schools.

Our team includes:

• Teacher Trainers: They understand how to effectively use technology in classrooms, ensuring Mindspark complements, not replaces, teachers.

• Field Staff – The Backbone of Implementation: This dedicated group is the heart of Mindspark's reach. They navigate the toughest terrains, carrying servers and tablets from village to village, ensuring no child is left behind due to lack of internet or electricity. Their proximity to the communities allows them to understand the specific challenges and contexts students face.

Mindspark isn't a solution designed for marginalised communities – it's a solution designed with them. We partner with both official and local entities, ensuring Mindspark aligns with government education goals while incorporating the insights of those closest to the communities. We actively seek input from communities. This ensures Mindspark addresses their specific needs and preferences, not a one-size-fits-all approach.

Mindspark is used by more than 500K students in India, UAE, South Africa, and Mozambique in public and private schools. In India, it is used in 10 languages in 15 states through three models.

1. Partnerships with 10 NGOs serve 195K students in rural, tribal and urban fringe areas of India, highlighting Ei’s commitment to collaborations, leveraging the expertise of NGOs to provide last-mile delivery.

2. Mindspark directly serves 207K students in government schools. Ei's staff work with education officials, oversee Mindspark sessions and support teachers. This approach is used for process discovery and government advocacy.

3. Mindspark serves 100K children at home or in community centers to ensure learning continuity beyond the schools (e.g. holidays, severe weather, political activities). It also provides an option for school dropouts and those forced to stay at home for family care, labor work or long funerals.

Mindspark's adoption in public schools prompted the Ministry of Education to issue a guidebook for Indian states, detailing PAL implementation strategies, with Mindspark as a notable case study. This demand, seen in large-scale procurement projects (e.g., Rajasthan bought computers for 5000 schools) and increased PAL contracts (e.g., Andhra Pradesh partnered with three EdTech providers). In research, Mindspark is recognized in reputed publications (e.g., American Economic Review, World Bank's "Smart Buys" report, Stanford Social Innovation Review).

MIT Solve's support can enable us to scale Mindspark, testing and making it available in countries facing similar or more severe learning challenges than India in the global south. This collaboration holds the promise of offering cost-effective solutions to enhance education outcomes worldwide.

Being part of the MIT Solve community offers us valuable opportunities:

1. Knowledge Exchange: Being part of the MIT Solve community enables us to exchange knowledge and ideas with other innovators and experts in the field of education technology. This exchange can lead to valuable insights and collaborations that can further enhance Mindspark's impact.

2. Building Evidence: We can gather substantial evidence on the improvement in learning resulting from home-based EdTech implementation, especially in remote, blended, at-home, and community usage models. This evidence will be crucial in demonstrating the effectiveness of Mindspark.

3. Process Discovery: Through documenting best practices with video and textual material, we can discover and share effective processes, enhancing the impact of our program.

4. Access to International Networks: MIT Solve provides access to international networks and partnerships, potentially making Mindspark accessible to more students globally.

5. Access to Global Funders: Being part of MIT Solve gives us access to funders at a global level. This access can help us expand Mindspark's reach and impact by taking it to more children around the world.

Mindspark's comprehensive approach to personalized learning, supported by pedagogical research and innovative features, sets it apart, leading to significant productivity gains and effective learning outcomes.

Based on research: Using pedagogical research and rich data from 20 years of assessments conducted in government and private schools, Mindspark captures specific misconceptions and helps overcome them proactively while enabling learning. 

J-PAL conducted a study on Mindspark in Rajasthan, involving 80 schools and 6,500 children over three years, with results presented at the RISE Conference 2019. The study found consistent positive results, with an ~80-100% increment in productivity in middle schools, regardless of starting ability level, gender, or grade.

In the QEI Development Impact Bond, which aimed to improve learning outcomes, students using Mindspark showed 7x more learning in the first year and 5x more over three years compared to control schools, despite the challenges of COVID-19.

Assessment for learning: Diagnostic tests in Mindspark detect misconceptions and proficiency in concepts and put students in personalized remedial paths that consist of questions and activities.

Real-life examples: To pique interest in learning a new concept taught, Mindspark facilitates learning by creating the need to learn the concept either by showing real-life situations and/or posing a problem that is efficiently solved if the targeted concept is understood. 

Intelligent responses: Mindspark has sophisticated and intelligent responses like response-specific feedback & alerts, numeric and non-numeric parsers to interpret various forms of student answers (like 'two' or '2' or '4/2' or '2/1' are the same)

Intelligent feedback: Step By Step solver in Algebra, Exponents, Real numbers and Trigonometry allows Mindspark to provide very specific and contextually relevant feedback messages, at each step. This granular data is further used to develop insights.

In addition, Mindspark English offers a variety of features such as Poetree, which encourages creative expression through poetry; Speaking, which allows students to practice speaking in multiple formats and genres; and Writing Tasks, where students can write essays on various topics or choose their own. It also provides Rubrics and spell-checks to assist teachers in evaluating student writing, and it offers a wide range of topics across grades, providing a comprehensive and personalized learning experience.

Games and enrichment activities: Mindspark has pedagogically oriented games that ensure the engagement of children and enrichment activities allow children who have learned a concept to explore beyond the typical curriculum.

Teacher interface: It ensures teachers have complete control over what concepts the class needs to learn, and how that can be effectively timed (in sync with classroom teaching). Further, teachers are provided with student and class level insights with recommended action points based on class performance. 

Ei’s vision with Mindspark is to use data on student learning and misconceptions to create an adaptive learning tool that can help each child learn at a level and pace most suited to them. Through an artificial intelligence algorithm honed over the years, the software helps remedy learning gaps and misconceptions through questions, activities, games, and challenges that are tailored to each user. Mindspark content is mapped to national policy and local context which can therefore meaningfully complement in-class instruction. Mindspark empowers teachers by allowing them to both control the activation of topics and use student data to understand progress and learning gaps. With real time automated data collection and analytics, Mindspark enables a transparent, easily accessible data dashboard to measure and track on the agreed M&E framework (on inputs, process, outputs & outcomes). This allows for effective monitoring and enables data-based decision making at both the micro (student and class) and macro (systemic) level. 

Ei has been integrating Machine Learning and AI into its products, particularly Mindspark, to enhance adaptivity, personalization, engagement, and feedback for learners. These technologies analyze student responses and interactions to tailor the learning experience. For adaptivity, ML algorithms adjust the difficulty level and pace of learning, ensuring each student receives content suited to their skill level. Personalization is achieved through the creation of customized learning paths based on individual progress and needs. This approach enhances engagement by providing real-time feedback, interactive exercises, and adaptive challenges, keeping students motivated. Additionally, the use of ML enables Mindspark to offer immediate and detailed feedback, helping students correct misconceptions and reinforce learning. Ei's goal is to further develop its expertise in expertise in ML and AI, including Generative AI to improve the teaching of foundational skills in a self-learning environment, focusing on refining adaptivity, personalization, engagement, and feedback mechanisms for enhanced learning outcomes.

Our solution team is strategically structured with 250 full-time staff, including AI specialists, software developers, content creators, and project managers, leading core development functions. We also have approximately 50 part-time staff, including educational consultants and research analysts, contributing to specific aspects of the project. Additionally, we engage over 600 contractors and workers in geographically diverse locations for localized roles such as Lab Incharges, Teacher Trainers, and Field Supervisors, ensuring comprehensive coverage across all project sites. 

1. With over 20 years of experience, Ei has collaborated with 10,000+ schools to administer 15 million assessments and assisted 15,000+ teachers for classroom remediation. It is also working with India’s largest board, CBSE to reform its Grade 10/12 school leaving exams. Ei’s work has been recognised by UNESCO, Harvard Business School, RTI International, and Stanford Innovation Review among others. For its EdTech tool, Ei was selected as one of 15 finalists in 2017 by the Qatar WISE.

At Ei, we recognize the rich diversity of India and are committed to mirroring this diversity within our team. Our recruitment strategy includes outreach to varied socio-economic, caste, and religious backgrounds, ensuring that our hiring practices are inclusive and representative of the communities we serve. For example currently, we have a team of 600+ members who are directly hired from the same communities which we aim to benefit.

We actively seek out partnerships with educational institutions and NGOs across different states to tap into a wide talent pool, which includes underrepresented groups in the edtech sector. We also implement sensitivity training that covers cultural, caste, and religious awareness to foster an understanding and respectful work environment.

To minimize barriers to opportunity, we offer scholarships and sponsorships for continuing education in edtech fields, specifically targeting individuals from lower socioeconomic statuses and marginalized communities. Flexible work policies accommodate various cultural and religious observances, ensuring a welcoming workplace for all team members.

Our inclusivity efforts are continuously monitored and adjusted through feedback from diversity audits and team surveys, making sure we not only comply with but exceed the standard practices for diversity and inclusion in the workplace

Mindspark has been used by a million unique students over the past decade. Annually it’s used by about 500,000+ students globally – with the vast majority in India. The low-income students study in 5,000+ public schools in 17 Indian states and use the learning software in 10 languages (English, Hindi, Urdu, Odia, Telugu, Gujarati, Marathi, Punjabi, Tamil and Kannada).

Currently, Mindspark program in government schools is broadly funded through a mix of corporate funding (70%), international grants (25%) and public procurement via government tenders. EI anticipates that short term and medium-term sustainability to be supported by hybrid funding models (philanthropy + government procurement) and wishes to set itself up for a more balanced funding base which will be stable and scalable.

An ideal division of expected funding sources would be an equal split between philanthropy and government initially with the latter increasing the % of its share. This mode of funding is critical to ensure an optimal buy-in from state and non-state stakeholders, and build consensus with practitioners, researchers, policy makers and administration. Early evidence of government procurement is seen in some Indian States like Andhra Pradesh, Haryana and a Social Impact Bond sponsored by Michael and Susan Dell Foundation in Uttar Pradesh. The World Bank and Asian Development Bank have also announced plans for 11 States to implement PAL.

While Ei is already financially viable, we are looking at the MIT SOLVE award to demonstrate PAL implementation in an initial set of 50 schools and provide guidance to create a blueprint for further scale-up to 5000+ schools in India and more globally. EI’s work on developing specific technical and financial specifications will aid the government capacity to evaluate and procure learning software. This will be an exemplar for state and international education departments to build capacity for public procurement, and implement effective structures that facilitate data driven decision-making. This will eventually intercept the low-accountability, low-learning equilibrium.

Ei also expects the current wave of educational technology adoption by the government, ushered in by the pandemic and emphasized in the National Education Policy 2020, to augment the focus and capacity to invest in digital blended learning. For scaling the innovation to 5,000+ schools, EI is exploring different pay-for-success models to be funded through philanthropic capital. The partnership is being anchored by Social Finance India which has the expertise in executing and securing such partnerships globally. A notable result-based financing scheme, STARS , launched by the World Bank in 6 Indian states, also ties disbursement to improved student proficiency, learning assessments and teacher development, with a special focus on the socially and economically marginalized. Asian Development Bank is considering a similar program for 5 more states.

Ei expects that MIT SOLVE award grant will create the needed momentum and salience for political and administrative support. This will enable the development of several coalitions that are invested in the future of 10 million disadvantaged children. These coalitions will drive the development dialogue in the country, with multilaterals, corporates, political leaders, governments, think-tanks, and policy makers shaping the future of education in India such as the MoE, NITI Aayog, the State Governments, and more. The government will be required to meet the implementation costs for the next phase of 8000+ schools, which can be met through the existing MoE’s ICT in Schools policy. States can apply for allocated per-school funding of $25,000 for setting up PAL labs. The states can re-apply for funding every 5 years by demonstrating learning gains. This amount will be sufficient to cover the costs for hardware and software deployment, teacher training and data dashboards for all stakeholders.

Mindspark currently costs about $30 per child/year inclusive of hardware and training, i.e. less than 5% of the current annual public expenditure per student. The government has already deployed computing hardware in over 50,000 schools over the past decade. The government will continue to progressively update the school infrastructure with computers, which can be used to run learning software in these schools.