MathDash

Learning is not supposed to be easy. It’s supposed to hurt your brain, the same way that your muscles must get sore in order to build new muscle and get more physically fit. The solution to improving learning is not to make learning “easier” (this makes learning worse) but instead it is to make learning more fun, more motivating, and more available to people.

MathDash is an online platform designed to transform the way mathematics is learned and practiced. It takes the fundamental concept of learning math and transforms it into a sport by incorporating competition, game mechanics, and social interaction. This unique approach is tailored to make math engaging and fun, fostering a community where users are motivated to improve their skills continuously.

MathDash is currently an interactive, web-based platform that allows users of all math skill levels to engage in competitive math challenges. These challenges are designed to match each user’s abilities and encourage progress through a well-structured system of divisions and ratings, similar to those found in competitive sports leagues.

Upon joining MathDash, users are evaluated to determine their math skills and are assigned an ELO rating and a corresponding division. Challenges presented to the users are tailored to their division, ensuring that the problems are neither too easy nor too difficult, which promotes a balanced and progressive learning environment.

Users compete against peers in their division, solving math problems to earn points and climb the ranks within their division. This competitive format not only makes learning more exciting but also drives users to consistently engage with the platform and improve their math skills.

After competitions, users can join discussion rooms to talk about the challenges, share solutions, and learn collectively. This community aspect fosters collaboration and deepens understanding.

MathDash leverages several technologies:

• Artificial Intelligence (AI): AI plays a crucial role in MathDash by assisting in the creation and review of math problems. It helps ensure that the challenges are of appropriate difficulty and quality. AI also performs automated answer checking to minimize errors and tags problems with relevant mathematical subjects, which aids in organizing the content effectively.
• Web Development: As an online platform, MathDash utilizes advanced web development technologies to provide a seamless and responsive user experience. This includes real-time updates, interactive features, and a user-friendly interface that accommodates a global user base.
• Social Tools: Integration of social tools allows users to communicate, discuss, and collaborate. Features like discussion rooms and leaderboards enhance the community feel and competitive spirit.

Please try out the site yourself at https://mathdash.live! A short gif can be found on the home page if you scroll down on a desktop.

MathDash makes learning mathematics more engaging, measureable, and accessible. This impact is especially crucial for young learners, including Black and Latino students and those experiencing poverty, who often lack equitable access to quality educational resources.

1. Customization and Accessibility:

MathDash utilizes AI to assess each user's ability level accurately and then assigns them to an appropriate division where the challenges are suitably challenging but not overwhelming. This personalized approach ensures that every child, regardless of their initial skill level or background, can start learning at their own pace, building confidence and skills progressively. The platform's design is intuitive and user-friendly, ensuring that even young learners can navigate it easily, reducing barriers to entry that might be technological or skill-based.

2. Encouraging Engagement through Gamification:

By turning math into a competitive sport, MathDash taps into the natural playfulness and competitiveness of children. This method is particularly effective in keeping students engaged and motivated to learn, which is crucial for priority learners who might not have strong educational support at home. The reward system based on points and rankings adds an extra layer of incentive for continuous learning and improvement. There is also an option to train on your own for students who prefer that.

3. Social Learning and Community Building:

MathDash fosters a sense of community among users, which is vital for learners from disadvantaged backgrounds. By participating in discussions about math problems and collaborating on solutions, students develop not only their math skills but also their social and communication skills. This peer interaction is especially beneficial in providing role models and encouragement from fellow learners who share similar experiences and challenges.

4. Enhancing Educator Resources:

For educators, MathDash provides a powerful tool to track the progress of their students through AI-driven analytics. Teachers can identify areas where students struggle and tailor their instructional methods or intervene more effectively. This capability is particularly valuable in diverse classrooms where educators face the challenge of meeting varied learning needs.

5. Scalability and Reach:

Given that MathDash is an online platform, it has the potential to reach a wide audience without the need for substantial resources that many low-income communities lack. Schools and community centers can implement it without the need for extensive infrastructure investment. This scalability ensures that the benefits of MathDash can be extended to a large number of priority learners across different regions.

6. Support for Continuous Learning:

MathDash can be played anywhere at any time, which is crucial for closing the learning gap for students who may not receive sufficient educational stimulation at home. By making learning resources available anytime, MathDash helps mitigate the impact of disruptions like those experienced during the COVID-19 pandemic, ensuring that learning can continue unimpeded.

In summary, MathDash leverages AI and an engaging platform design to meet the specific needs of priority learners by making math learning accessible, personalized, and enjoyable. This approach not only enhances educational outcomes for these students but also supports educators in delivering more effective and differentiated instruction.

We grew up on math games, and know just how powerful their effect is on providing both learning opportunities and motivation.

Akshaj and I met when we were just 11 years old in 2011, playing an online math game called For The Win! The game was extremely simple - two players would race to solve a single math problem. The first to get it would receive one point, and the first to four points would win.

We played this game and others for thousands of hours in our childhood because it was the most fun way to engage with math. This was my video game - in part because I thought it was good for me, but also in part because it was genuinely fun. The games had a rating attached to them, and I wanted my rating to go up, and I wanted to win.

As a result, we got exposure to hundreds of math concepts we never would have learned otherwise, and got extremely good at math. We participated in all the competition math circuits that were currently available, and were multiple-time USA Math Olympians. Akshaj competed on ESPN at a competition called MathCounts. We were both invited by the US government to the elite math students training camp called the Math Olympiad Summer Program, for the top ~50 students in the nation, based on the USAMO (https://en.wikipedia.org/wiki/Mathematical_Olympiad_Program). 

There we became lifelong friends and also grew an interest in education, becoming president of our high school math team and coaching other local math groups. I personally taught math to not just extremely advanced students, but also in low income areas and advocated to run math programs for people who did not have exposure to math yet because I started out in a low income area and only discovered math because I stumbled upon a math competition which led me to the math game, so I understand the needs of the general community as well (not just the elite math students). We taught the official web development class at MIT, and are currently running MathDash which already has 500 daily active users we chat with every day in our Discord.

Our users excitedly chat with us every day about what feature improvements they want, and discuss math problems with each other. We build them ASAP since we are some of the fastest web engineers. We are basically building for past versions of ourselves and also for students who we coached. It’s a burgeoning community and we are in the center of it!

We have a website with 500 daily active users and 1500 weekly active users with 40% one month cohort retention and 20% week over week organic growth.

We have a product that people enjoy playing and have a highly retained audience, but we want to have a little more conviction that it has broad appeal before entering the high growth stage.  This audience is mostly niche math competition people though we are starting to see signs of people with no exposure to contest math engaging with our games.

New York

MathDash represents a significant innovation in the educational landscape by transforming the way mathematics is learned and engaged with, especially for Pre-K-8 students. Our solution is not merely a digital tool for learning math; it reimagines math as a dynamic, competitive sport, making it inherently more engaging and accessible for students of all skill levels.

Innovative Aspects of MathDash:

1. Gamification of Learning: Creating a Novel Game: MathDash increases engagement and motivation through creating an original game. People want to get better at the game because it's fun, and as a result pick up problem solving skills as a byproduct. This sports-like format is familiar and exciting to students, which helps reduce anxiety associated with traditional math learning and encourages a more enthusiastic engagement with mathematical concepts.

2. Personalized Learning Paths: Utilizing an ELO rating system typically found in chess, MathDash assigns users to divisions based on their math ability, ensuring that challenges are appropriately difficult for their skill level. This personalization makes learning adaptive and tailored, vastly improving the effectiveness of educational content delivery compared to one-size-fits-all approaches.

3. AI-Enhanced Problem Creation and Assessment: The integration of AI not only ensures the scalability of creating and assessing math problems but also enhances the accuracy and relevance of these problems. Our AI tools assist in brainstorming problem ideas, checking answers, and tagging problems with relevant subjects, which streamlines the organization and delivery of educational content.

4. Community and Collaboration: MathDash leverages the power of community by enabling students to collaborate and compete with each other. This social aspect not only makes learning more enjoyable but also builds a supportive network where students can share knowledge and strategies, enhancing peer learning and emotional support.

5. Empowerment through User-Created Content: Encouraging users to contribute to problem creation democratizes the learning process and empowers students by allowing them to shape their learning environment. This feature also fosters a deeper understanding and appreciation of mathematics as students engage creatively with content creation.

MathDash's approach could catalyze significant changes in the broader educational market by demonstrating the effectiveness of gamified learning environments and competitive frameworks in educational settings. By proving that competitive, gamified learning can be both effective and scalable, MathDash has the potential to shift the educational focus from traditional, passive learning models to more active, engagement-based models. This shift could influence educational policy and resource allocation, prioritize interactive and student-centered learning environments, and encourage the development of similar platforms across various subjects and educational levels.

In conclusion, MathDash is not just an educational tool but a revolutionary approach that combines technology, psychology, and pedagogy to transform how students interact with math. By making math a competitive and collaborative sport, MathDash is poised to change the educational landscape, making learning more engaging, accessible, and effective for all students.

1. Natural Language Processing (NLP) and Machine Learning (ML):

• ChatGPT and Pre-Prompting: We integrate a customized version of ChatGPT, which utilizes NLP to assist in generating problem ideas. This AI-driven process helps in brainstorming diverse and creative math problems, ensuring that content remains fresh and engaging. Additionally, the NLP component is crucial for labeling and categorizing these problems based on their mathematical concepts and difficulty levels, streamlining the user experience.
• Answer Checking: NLP is also employed to interpret and verify users' answers against the correct solutions. This not only ensures accuracy in scoring but also provides immediate feedback to learners, an essential feature for effective learning.
• Recommender System: We are developing a recommender algorithm using machine learning techniques to analyze users' performance data, preferences, and learning patterns. This system will predict and suggest math problems that are optimally challenging for users, enhancing the personalized learning experience.

2. Deep Learning (DL):

• While currently not explicitly stated, the potential to incorporate deep learning exists, especially in improving the efficiency of our NLP models for problem generation and answer validation. This would involve training custom models on large datasets of educational content to refine accuracy and context-awareness.

1. Real-Time Competitive Framework:

• Synchronous Online Competitions: We leverage real-time web technologies to host live math competitions similar to platforms like Kahoot. This involves complex server-side logic to manage live interactions, score tracking, and dynamic updating of user standings.
• ELO Rating System: Inspired by chess rankings, the ELO system is adapted to continuously evaluate and update each user’s mathematical proficiency based on competition outcomes. This adaptive rating system helps in maintaining fair and motivating challenges within the user community.

2. Web Development Technologies:

• Frontend Development: The user interface of MathDash is built using modern web development frameworks like React.js, which allows for dynamic, responsive, and interactive user experiences. This choice supports real-time updates and interactive functionalities essential for hosting live competitions and user interactions.
• Backend and Database Management: For backend operations, we use Node.js combined with a MongoDB database. This combination efficiently handles data transactions, user management, and real-time communication required for synchronous competitions and user progress tracking.

MathDash’s AI functionalities currently rely on third-party models like OpenAI’s ChatGPT for certain NLP tasks. While these models provide robust, state-of-the-art capabilities, our long-term goal includes developing in-house AI models tailored specifically to educational content and user interaction within MathDash. This will allow greater control over model behavior and customization based on user feedback and evolving educational needs.

In conclusion, MathDash integrates multiple AI technologies and advanced web development practices to create an innovative and dynamic platform for math learning. This blend of AI with real-time interactive technologies not only enhances the learning experience but also sets the stage for future innovations in educational technology.

We do not have much model testing/third part evaluation to offer, but the effect will be evident if you poke around our Discord for 30 seconds.

Anecdotally, we grew up in an early age of internet games, where users could start interacting online. From 6-8th grade, we were addicted to a math racing game called For the Win! created by the Art of Problem Solving. We actually met online in 6th grade from opposite coasts of the US because we were power users of this game and it got us addicted to math. I would strongly argue that this effect is causal - the thousands of hours we spent playing this math racing game helped us train our mental fitness and motivated us to learn a lot of math concepts.

Since then, we became USA Math Olympiad champions, got into MIT, worked at Jane Street, and also did YC. Our peers who played this game are similarly successful, including the founders of many unicorn companies. This approach worked for us, it worked for our peers, and it will work for everyone if introduced correctly.

MathDash is deeply committed to ensuring equity and combating bias in its AI implementation, particularly to serve priority learners including Black and Latino students and those experiencing poverty. Here are our key strategies and actions to ensure our AI technology promotes equity and fairness:

Our data sets are about math, which is a universal language. We are working on translations for some of the wordier problems into other languages!

MathDash problems are all screened and approved by a committee to detect and address any biases or unfair advantages that might arise in AI-generated content or user interactions. We analyze performance and engagement metrics across different demographic groups to identify and rectify disparities.

We collaborate with educators and experts from diverse educational settings to ensure that our content and algorithms are culturally relevant and sensitive. For example, we are working with Arun Shanmuganathan to bring our resources to the Pan African Mathematics Olympiads (PAMO) community.

MathDash is committed to transparency in how our AI functions and makes decisions. We provide clear explanations for how data is used, how models operate, and how decisions are made, which is crucial for trust and accountability. We also maintain an open channel through Discord for feedback on AI performance and are prepared to make adjustments based on community input. Our average response time is under 5 minutes!

We adhere to ethical AI practices by following guidelines that prevent discrimination and protect user privacy. Our AI development team is trained in ethical AI use, ensuring they are aware of and actively working to prevent potential biases.

By implementing these strategies, MathDash aims to create an equitable learning environment that recognizes and addresses the unique challenges faced by marginalized groups. Our goal is to ensure that every learner, regardless of their background or circumstances, has equal access to opportunities for learning and advancement through our platform.

Full time: 2
Intern: 1

We have already run contests with math circles and other math organizations around the world, including the Carnegie Mellon Informatics and Mathematics Competition, Stanford Mathematics Tournament, AlphaStar Math Academy, Minnesota Math Circle, and are working closely with Singapore Math and the Art Of Problem Solvingto spread their curriculum on our platform in a game-format.

We will operate on a freemium model, which means the basic solutions and access to games is free to all (currently everything is free).

We think this is the solution to and future for education. Education is supposed to be hard because learning is working out your brain. The solution to make working out fun was sports - we are creating a sport for education starting with math.

The Bill and Melinda Gates Foundation has a lot of influence and reach so we want to partner to get our name out there and have more people competitively learning.