ERI: Educational Real time Interface

ERI can deliver free private tutoring powered by peer-to-peer learning and artificial intelligence augmentation.

Across the world, more than 1.8 billion students lack access to education beyond the equivalent of the fifth grade. It’s been shown that students who regularly receive private tutoring are able to score two standard deviations above students who do not on standardized testing. That said, the demand for private tutoring far outstrips the supply, with up to 65% of students seeking sessions in Kenya and 73%  in Sri Lanka. Language and mathematical literacy is the backbone of the modern economy. In the U.S. alone, more than 36 million adults lack basic literacy skills, costing an estimated $225 billion in productivity and tax revenue. Private tutoring has been proven to be the most effective way to improve learning outcomes, overall literacy and as a consequence, increase graduation rates. It’s clear that there are multiple problems: access to tutoring, the quality level of tutoring and the high cost of 1-on-1 sessions.  

Our proposed solution endeavors to mitigate all three by using artificial intelligence with peer-to-peer tutoring. ERI (Educational Real-time Interface), is a tutoring platform that is free for students. ERI has three main components: a mapper to identify and build a knowledge map of the students’ skills, a matcher to pair students according to their needs and the amplifier which elevates the quality of the tutoring to a professional level by suggesting AI-generated responses for the peer-tutor. Moreover, 21st century skills such as communication, collaboration and problem solving can be actively enhanced by allowing students to learn from one another in a peer-to-peer context.

Not only can students can help each other with the majority of difficulties experienced while solving math problems, but also our platform will output a report to the teacher with a breakdown of which exercises or types of difficulties were experienced and would require more attention.

Indeed, teachers spend a large portion of their time correcting homework (procedural, or conceptual errors). While procedural mistakes are very common, they are very simple to fix and can be accomplished in a peer-to-peer context so that teachers can spend more time on conceptual mistakes that are more pedagogically important. Our solution can improve and shorten homework correction time without a negative effect on learning outcomes. It enables teachers to delegate the correction of procedural math homework mistakes to the students themselves with assistance from artificial intelligence. Thus, teachers have more time for meaningful tasks in the classroom.

We strongly believe that access to education provides viable skills and valuable knowledge. We want to ensure that every human being has that access to a quality education —leading to productive existence for all. We see great potential for AI technology to democratize expertise and knowledge for many individuals worldwide. We are not looking to replace humans with machines, instead we seek to augment and amplify what humans can do through Artificial Intelligence with our platform, ERI.

By combining human knowledge and AI innovation, ERI aims to provide a quality equivalent to professional tutoring — but affordable and accessible everywhere at anytime.

The novelty of our solution centers around a tight interaction between users and a generative model informed by an understanding of the problem, the conversational context, and the dialogues of other tutors, to empower users to create higher-quality interactions. While classic techniques (using templated answers or prior knowledge to build rule-based systems) for writing bots by hand have succeeded in closed domain settings, our techniques are designed with the more complex interactions inherent in tutoring.

ERI’s system operates based on the performance of the interactions. It is customizable and adaptive to ongoing conversations where prompts shown to the peer tutor to enable them improve their communication and knowledge transfer skills. We ensure the system is easy to use and handy for students to engage in a conversation with peers. We plan to validate the impact of the technology with ANOVA statistical analysis and regular surveys to gather additional key insights. We will integrate an off-the-shelf analytics technology to analyze system usability with heatmaps and mouse tracking to ensure that teachers understand and extract quality information.

We are currently running a pilot version of ERI and aim to release the public version in September 2018. Until December 2018, we will analyze the findings and improve the overall system quality.

For 2019, we are seeking to expand our platform with a powerful analytics module for teachers. It will provide the teachers key metrics in an easy to understand dashboard so that they can prioritize their lessons based on the problems identified in the peer-to-peer interactions. We will wrangle the collected data and extract relative information about the features that we could improve or include in future versions.

Our goal is to reach a million students worldwide within 5 years. We also want to cover different topics than math (literacy, foreign languages, science…) and provide open source educational resources within the platform, to achieve the best user experience for both teachers and students.

Our ultimate goal is that education becomes accessible for free to any student wherever they are on the planet. By building greater collaboration between students, we believe that our AI-empowered peer-to-peer system empowers students for success and a better life.

The fact that our solution is free is definitely appealing to people all around the world. We plan on using our current network in education (schools and NGOs) as well as media coverage to spread the word about ERI. Then the quality and the accessibility of the product shall retain customers. Besides, ERI is available on computers as well as on mobile devices so that students even without Internet connection can use the phone lines to communicate wherever they are.

Our solution is currently being beta tested with 1,000 high school and middle school students participating in two different regions and populations of the world - Montreal, Canada and Mathare, Kenya; in French and in English. We are currently collecting data and feedback to improve both UI/UX and to continually train the AI.

In the past few months, we have conducted research which evaluated the effectiveness of an AI-empowered dialogue recommender with positive and robust results. ERI meets two United Nations sustainable development goals: Quality Education and Decent Work and Economic Growth.

The next year, we plan to expand organically to 30,000 users in areas where French and English are widely spoken: Africa, the Caribbean, India as well as the USA.

Then within 3 years, we want to reach 500k users and have our platform in refugee camps as they are the places where access to education is severely limited. We would also begin our expansion into other disciplines outside of mathematics.

Our team is comprised of 6 individuals who amongst them have founded a total of 4 startups, 20 years of AI expertise, and 20 years of domain expertise. Members have different backgrounds: technology, math, neuroscience, social sciences, and education. This diversity of perspectives enables us to continuously innovate as a brand; bouncing ideas off each other,  individually giving a unique point of view. Everyone is very committed to our mission of helping students and educators all over the world through cutting edge technology.

Our revenue will be generated through SaaS model with educational organizations (private tutoring companies, libraries, universities...) through the licensing of our AI technology developed and tested in our free market segment. We provide a custom version of ERI (or of its AI bricks) to private companies which want to implement a specific AI-messaging system within their existing solutions.

We are applying to SOLVE because our solution addresses the main point of the Teachers and Educators challenge, which is to amplify and support through technology to offer quality education. We believe that Artificial Intelligence can greatly benefit society but at our core, we believe that it should augment and collaborate with humans, instead of replacing them.

For our AI to be efficient it needs to be fueled by a large amount of relevant data; the more students we have, the better the solution becomes. So the development time depends a lot on the number of onboarded users and their utilization of ERI.

By providing relevant partnerships, Solve would greatly help our development and impact.