SBIR-STTR Award

This SBIR Phase I project would result in the development of an intelligent automated tutoring system for teaching coding to K-12 students. Millions of 6-14 year old students in the U.S. have an interest in learning how to mod ("modify") popular video games, but lack the necessary skillset: coding. Coding is a highly valuable STEM subject, but is not currently a core subject in U.S. schools. Because the U.S. is not producing enough skilled workers in computing fields, there will be 56,000 jobs that go unfilled by U.S. workers every year. This project aims to expose millions of students to coding in the next 5 years by making modding accessible to them. In order to teach millions of students at scale, this project will develop a complex automated tutoring system which intelligently guides students through their coding education, detects and helps students when they run into errors, and provides students with a near-endless supply of puzzles and educational content. This project plans to tackle ongoing problems in the field of computer science education. The automated tutoring system developed through this project will be a foundation for automated tutoring systems that teach other STEM subjects.

The main technical innovation of this project is a browser-based IDE for writing code and executing it within the runtime environment of video games. This inherently gamified coding environment opens new doors in computer science education. Around this coding environment, an automated tutoring system is being built. This project aims to tackle several problems in computer science education. These problems include: (1) a stealth assessment system that identifies when a student does not understand a coding concept and that guides the student towards understanding, (2) a novice-friendly error detection and reporting system that points students to the part of their code where the error originated, (3) an automated Stack Overflow-like resource for novice programmers that suggests possible solutions when the novice encounters an error, (4) a random puzzle generator that creates a near-endless supply of coding educational content for different skill levels, (5) tutorials and puzzles that gamify the transition from a visual programming language to a syntactically complex programming language, and (6) teacher tools and APIs that allow teachers to integrate the software into their classrooms. By solving these problems in computer science education, this project aims to make an effective coding education tool that scales.

This Phase II project proposes to develop a computer science (CS) educational software that has the potential to inspire millions of U.S. K-12 students to learn computer programming. This software will leverage the motivational power of a popular video game, to teach CS to students by teaching them to reprogram the video game itself. The United States currently has a severe deficit of students pursuing CS. The Bureau of Labor Statistics predicts that over 1 million computing job openings will go unfilled by U.S. workers by 2022. By leveraging the power of a popular video game, the technology proposed in this Phase II project has the potential to expose millions of K-12 students to coding in the next 5 years. The commercial impact of the underlying technology developed in this Phase II project does not stop at the over 100 million users who currently play the popular video game with which the current educational software integrates. Because the underlying technology is transferable to any moddable (i.e. reprogrammable) video game, the technology has the potential to be used to teach CS with other popular titles from the rapidly growing video game industry.This Phase II project proposes to continue the development of a software product that is a web-based coding environment for novice programmers. This software goes beyond the state-of-the-art technologies in this space (i.e., scratch.mit.edu) in several ways: 1) It uses automated tutoring techniques to customize the educational experiences for novices, 2) it facilitates writing programs that manipulate objects and terrain in a 3D environment, 3) it allows the novice user to reprogram a popular video game, 4) it has an in-browser, WebGL-based 3D runtime environment, 5) it supports both a novice-friendly visual programming language (Blockly) as well as a text-based language (JavaScript), 6) it leverages gamification techniques such as badges, points, and unlockable items, and 7) it supports multi-user, collaborative coding. The objectives of this Phase II project concentrate on improving student experiences in order to increase customer retention and acquisition and to finish the development of a marketable product that will teach 5 million K-12 students in the next 5 years. The first objective of this project involves developing and extending the browser-embedded game engine. The second objective focuses on improving systems that match students with appropriate educational content and motivate students to continue learning. Finally, the third objective involves implementing new systems that incentivize students to create and share with the community.