EIDU

EIDU is committed to solving the global learning crisis. Many educational interventions and learning solutions have been implemented over the years, with some achieving high impact. Yet on their own these solutions do not solve the global learning crisis. Instead, if we integrate all these thousands of educational solutions in a compatible and effective way, we can actually eradicate learning poverty. At EIDU, We have developed an open learning platform that has already scaled to 35.000 children at a cost of $3 per child per year. Our system integrates, assesses, and iterates on learning solutions at high speed. We built an open-plugin system where content creators, assessment developers, and AI-optimisation builders can all integrate and test their solutions. By generating more learning data in 1 day than PISA does every 3 years, we can quantify the exact impact of every intervention we make, big or small, revolutionising global education. 

More than 600 million children worldwide are not receiving quality education, even when attending school. In high income countries, 10% of the children fail to reach minimum proficiency levels in reading and in math. In low income countries, up to 90% of the children do not reach these same minimum proficiency levels (https://en.unesco.org/gem-report/report/2017/accountability-education, https://en.unesco.org/gem-report/allreports) . This exemplifies the huge inequality in opportunity for children to fulfil their potential. The root-causes of this inequality are multi-faceted (https://www.worldbank.org/en/publication/wdr2018 , https://documents1.worldbank.o...); teachers lacking the didactic and pedagogical skills to teach and assess, poor quality of educational resources, children poorly prepared for learning, ineffective school leadership, and a lack of secure learning data supporting governments in rolling out effective policies and accountability structures. Most interventions target only one or a few of these issues. Moreover, these  approaches tend to be small-scale and developed too slowly. We need a system that brings together all these learning solutions in a meaningful way, and allows for innovation in the educational space to happen at a pace which is greater than anything we’ve seen before.

Our standardised interface ensures resources such as digital exercises for learners, teaching plans for teachers or administrative overviews for management can quickly be integrated onto a single platform. Our validated digitised assessment tools allow us to measure the impact of every piece of content in isolation by continuous A/B testing. We currently collect 4,000 assessment results every single day, allowing us to detect even marginal learning gains within our vast dataset. How and when this content should be presented to each individual user is optimised by allowing various personalisation algorithms to compete against each other within a multi-armed bandit framework, optimising on learning impact. On top of this automated process, we constantly analyse our results validating, adapting and iterating in a virtuous cycle of improvement. Vitally, experts around the world can contribute at any stage of this process from content-creators to academics to machine learning gurus since we make this whole process open through plug-ins. Because only together can we overcome a challenge of this scale.  

In Subsaharan Africa alone, over 200 million children are not reaching minimum proficiency levels and these numbers are forecasted to remain static over the coming years. This is despite extensive commitments related to SDG 4 (inclusive and equitable quality education for all) from around the world. The situation is particularly dire in urban informal settlements which normally house the poorest communities. Here only about 10% of children who are attending primary schools reach minimum proficiency levels.

 In order to build up our understanding of the communities we want to serve, we set up a user lab in one of Africa’s biggest urban informal settlements, Kibera (Nairobi, Kenya) in 2016. By continuously seeking and incorporating feedback, our earliest users have helped us develop a platform which is adapted to these environments, rather than simply importing solutions that are out-of-context. This means the platform is capable of delivering impact in challenging circumstances and is valued by the teachers and schools that use us.

 Once a school has signed up for EIDU, each teacher gets a smartphone with the EIDU platform installed and ready for her and her learners to use in the classroom. They don’t need internet access to use the platform, even an intermittent connection is sufficient to upload usage data and download updates.

 Our tailored digital environment is further complemented by our coach-system. Each school receives continuous, context-driven, interpersonal support from other members of their community. This also allows for much more open channels of communication from our users to help inform us how the platform should develop.

 The success of our implementation model and the support of local communities for our initiative means the EIDU platform is now actively used in thousands of low-resource schools in Kenya (Nairobi, Mombasa) and Nigeria (Lagos), with an active daily student population of 35000.

 Our students are engaging with the content from x-prize winner “onebillion.org”, whose content has been proven to result in significant learning gains in various studies (https://www.imagineworldwide.org/wp-content/uploads/An-8-month-RTC-in-Malawi_Final-Report_Jan-2020.pdf, https://pubmed.ncbi.nlm.nih.gov/31120168/). This is in addition to “anton.app”, one of Europe’s most widely used learning apps; helping all students of the class to develop their early literacy and mathematics competencies.

 Each child works through the content at their own pace, maximising their opportunity for learning. With the integration and constant improvement of AI-personalisation algorithms, students will engage with more and more appropriate exercises and reach ever higher levels of learning. In addition to our personalised learning content, our system will be integrating structured pedagogy programmes, such as teacher guides and coaching support. This allows for teachers to develop themselves professionally and support their students much more effectively. This teacher focused content will also be continuously validated and personalised to ensure maximum impact. Only by engaging with every stakeholder in a child’s education and by actively seeking to understand their needs can we hope to bridge the global education gap.

Our platform is specifically designed to work in the most challenging of environments. Our students are from low-resource schools in African urban centres which historically have the lowest attainment. We provide a diverse set of content to ensure the engagement of our learners. Our continuous testing and iteration facilitates children to learn effectively and independently. We are branching out beyond learner content into every aspect of education from teacher guides to parental support. By delivering the best possible learning resources in a holistic, easy-to-use format we strengthen the whole education ecosystem, giving these children a real chance of success.

EIDU has spent the last 5 years perfecting our implementation model and building our learning platform. Currently, over 35.000 students in over 2.000 low-resource schools are working with the platform every day, based in Kenya (Nairobi, Mombasa), and Nigeria (Lagos). We secured funding to roll out our model to over 250.000 students in the next 2 years, adding Ghana to our user base, and are looking for a next round of funding to expand beyond that.

So far nobody in the world has managed to scale a technology-based solution to low-resource educational settings, especially not at $3 per child per year. Our innovative implementation model shows that we can reach all children, in all educational settings of the world. 

So far, nobody in the world has managed to gather accurate, real-time, continuous learning data of marginalised children from low-resource environments. By gathering over 4.000 test scores per day, we are absolutely unique in understanding what learning is happening and how it is happening. On top of this our reach and capabilities are quickly growing.

So far, nobody has analysed learning solutions for low-resource environments to the smallest detail, allowing for iterative improvements, and gathering real understanding of what works in these environments.

We currently have 35,000 students and 2,000 teachers actively using the platform. An additional user is 1 county government with whom we're running a pilot to test our public sector collaboration.

Next year we're estimating to serve up to 200,000 students, 6,000 teachers, and 1 county government

In 5 years from now, we're estimating to serve up to 10,000,000 students, 30,000 teachers, and 4 national governments

Using internationally validated and standardised assessment batteries that have been used in over 26 LMIC (EGMA https://shared.rti.org/content... and EGRA https://shared.rti.org/content...) we measure children's learning progress daily, targeting the most predictive learning constructs as selected for the EGMA and EGRA, focusing on early literacy and early numeracy. In the next few years we will be adding several more internationally recognised assessment batteries to further deepen our ability to assess children's learning, both from preschool age up to end of primary school age.

35 people, all fulltime

We have a dedicated team https://eidu.com/team/ with extensive experience in successful digital businesses, product development, data science, and educational science.

Our leadership team are very experiences software entrepreneurs, having built and scaled companies that are valued about $100 million.

Our dedicated field team understands our user base to the core, having been teachers in low-resource schools themselves.

Our product team is highly experienced in product design, development, and implementation.

Our data expert has an Msc. in mathematics and has worked in various places engaging with AI technology.

Our learning expert has an Msc. in child psychology and pedagogical sciences.

The vast majority of our staff has been with the company from the earliest years onward and is there to stay, helping EIDU reach it's mission to solve the global learning crisis.

Since we are building a global company, with teams both in Africa as well as in Europe, we are committed to establishing a community that celebrates diversity and inclusivity. From day one we have been a culturally diverse team with men and women from very diverse backgrounds, and from day  one we've been witness to the elevated level of discussion and understanding that is achieved when working with such a diverse crowd. If we want to be of use to our schools and children, we need to understand them, and we can only understand them, if we are them. 

The way our company has been built from the beginning means that 70% of our staff works and lives in Africa, 30% in Europe. Half of our employees are men, half are women. Some of us have enjoyed privileged upbringing, some have not.

We are committed in keeping our high standards of inclusivity, equity, and diversity. As we are growing, all of our staff members are growing with us, providing everyone a fair chance of professional development. Specifically within our hiring process we have learned and are still learning how to make sure everyone gets a fair chance, and to make sure that our celebrated diversity maintains one of our top priorities.

We're applying mainly for the connection to your amazing scientific community. We are in the process of fundraising, so the funding is of course meaningful. But most importantly, we are introducing AI technology into our system and we're eager to share and collaborate. As we have made clear throughout this application, the staggering scale of the learning crisis makes it too large for any one organisation to solve. Only by collaborating with thought leaders, especially in the technological space can we hope to provide solutions that end learning poverty. We believe becoming a solver means we enrich and become part of a network that is essential to this solution. 

Our whole system is based on solid M&E strategies and accurate learning data. We must reach the highest levels of certainty in knowing that the learning data that we're collecting is accurately showcasing children's development in basic literacy and numeracy skills. At the same time, we are collecting all learning data entirely digitally. Which is unique within our population. So by introducing an innovative approach to learning assessment, we seek active collaboration with experts in educational M&E.

Secondly, as mentioned before, we are introducing personalisation strategies into the system. This is still a fairly new field and there is still a large amount of innovation and research to come. We believe our data and our platform may well prove fundamental in bringing these innovations forward. By spreading our message and seeking active collaboration with the greatest minds we can success in this.

We are mostly interested in the scientific community engaging with AI for good causes. We're specifically interested in the "AI for humanity" prize.

EIDU uses strong data science and artificial intelligence to benefit humanity making it an ideal candidate for this prize. Despite it being fundamental to society, it is still surprisingly unclear how we can effectively measure and evaluate improvements in education. In an attempt to answer this, EIDU has created validated digitised assessment tools which are randomly assigned to learners. This means in addition to the data from the 500,000 exercises currently completed each day, we also get verified learning data from around 4,000 items. We randomise approximately 5% of our content allowing us to evaluate every piece of content in addition to all our classic A/B experiments. The only way to gain insight from this vast amount of complex data is of course a varied and thorough data science approach. Our AI application becomes relevant when we try to personalise all of this content for each individual child. To do this we have implemented a number of personalisation algorithms including Deep Knowledge Tracing, our own adaptation of ELO as well as various types of path selection informed by classification models. We are keenly aware however that our best chance of finding the most effective AI implementation is by opening our platform up to the world. We are therefore building infrastructure which will allow us to do this. Soon any team or individual in the world will be able to train models on our data. This prize will help us find the experts who want to contribute to solving this crisis.

EIDU uses strong data science and artificial intelligence to benefit humanity making it an ideal candidate for this prize. Despite it being fundamental to society, it is still surprisingly unclear how we can effectively measure and evaluate improvements in education. In an attempt to answer this, EIDU has created validated digitised assessment tools (EGMA, EGRA) which are randomly assigned to learners. This means in addition to the data from the 500,000 or so exercises currently completed each day, we also get verified learning data from around 4,000 items. We randomise approximately 5% of our content allowing us to evaluate every piece of content in addition to all our classic A/B experiments. The only way to gain insight from this vast amount of complex data is of course a varied and thorough data science approach. Our AI application becomes relevant when we try to personalise all of this content for each individual child. To do this we have implemented a number of personalisation algorithms including Deep Knowledge Tracing, our own adaptation of ELO as well as various types of regression analysis. We are keenly aware however that our best chance of finding the most effective AI implementation is by opening up our platform to the world. We are therefore building infrastructure which will allow us to do this. Soon any team or individual in the world will be able to train models on our data. This prize will help us find the experts who want to contribute to solving this crisis.