In today’s knowledge economy, education is the essential enabler for employment and opportunity. This is becoming especially urgent as millions of jobs are being replaced by automation and artificial intelligence, and in Western Pennsylvania we are ideally situated to understand the complexities of these economic shifts. Yet the existing education infrastructure today is already overburdened, with challenges surrounding uniform access, and tripling its scope to include people of all ages would be an impossible task.
Expii solves the universal training and learning problem through a novel self-organizing platform which unites the creativity of the world to deliver on-demand and personalized education for free. Expii leverages the massive penetration of smartphones to reach areas throughout the world, starting from its local research base just outside Carnegie Mellon University. The key insight is that in a modern world where content is ubiquitous and freely shared by the crowd (but of varying quality, and disorganized), the central question is then to identify exactly which piece of content a particular learner should interact with at any given moment, based on the learner’s current knowledge base and the learner’s long-term goals. Expii introduces a new mathematical framework for answering this question, which recasts the general problem of Education as a graph algorithm problem, informed by data collected from the universe of users across the platform. This new formulation also unlocks a paradigm shift away from courses and grades, towards a new ecosystem of atomic concepts and population-wide statistical micro-certifications.
The platform is released for free on expii.com (and as the free Expii app in the Google Play Store) with an innovative user experience that parallels a modern Maps app: learners indicate what area they would like to improve in at the moment, and they are provided with relevant problems (which lead to lessons) based on their profile of previous interactions with the site, dynamically calculated in real time.
Simultaneously, the system learns characteristics about a body of crowdsourced and openly licensed educational content (also collected by the platform). This dynamic system for automatically rating users and problems has far-reaching implications across a unified learning platform. Within each topic it guides users to problems of most appropriate difficulty, seamlessly moving to more challenging problems as users demonstrate proficiency, going beyond basic literacy with unlimited growth potential. The system will provide a complete adaptive educational solution to any person with a smartphone. Long-term, a social layer will connect learners with others who have high ratings as measured by the platform, scaling the human-to-human learning experience itself.
Expii seeks to further develop its platform in collaboration with community organizations in Western Pennsylvania, which has the diversity to represent a broad swathe of the United States population, with both urban and rural populations in close proximity. Its platform is under constant revision in response to feedback, and lessons learned in this region will scale into worldwide impact through new features in the globally available free product. Expii has already connected with several local educational units (as it has a collaborative attitude with the existing educational infrastructure), and aims to work with the region’s rich ecosystem of community organizations which operate on the front line. It is an active participant in the region’s Remake Learning Network, as well as in the local startup community. More globally, Expii is already in collaboration with communities from New York City to Bangladesh, and has collaborated with The New York Times and FiveThirtyEight (part of ESPN).
The problem decomposition and solution design are directed by founder Po-Shen Loh’s mathematical background as a Carnegie Mellon math professor, National Science Foundation CAREER award-winning researcher in network theory, probability, and theoretical computer science, and the National Coach of the USA Mathematical Olympiad team.
The core engine of Expii’s novel approach is based on the Elo Rating System, which has been used for half a century in the context of competitive Chess, where it identifies World Chess Champions and helps to arrange matches between players of comparable ability. This system has been applied to many contexts outside of Chess, such as online gaming, computer programming competitions, football, and more, but has not been applied at scale in Education. On Expii, each user has 10,000 ratings (one per concept in the Expii universe), and each problem has a rating and one (or more) concept tags. After each user-problem interaction, the problem’s rating and the user’s concept rating are both updated in real time based upon the outcome (success or failure), similarly to how chess players’ ratings are updated after each game. The ratings become a new metric for a user’s STEM ability (capable of measuring both conceptual and mechanical skills). Initially, ratings will be useful self-assessment tools; eventually, they will provide rich alternatives to existing metrics such as grades and standardized exam scores.
The platform itself is inspired by modern Internet products: it combines elements of Wikipedia and Quora, with content creation made as easy as Tumblr and Google Forms, organized like Khan Academy, but with the turn-by-turn navigational interface and browsability of Google Maps. It has the potential to organize and widely disseminate the largest body of openly-licensed educational resources ever created, into the hands of billions of people worldwide through their smartphones, tablets, and computers. In the near future, a social layer will be implemented, using the ratings to match learners with peers who can provide them with optimal help, thereby solving the student-to-teacher ratio problem by converting peers into teachers whenever there is potential for productive exchange of expertise.