Grey-Box: Enhancing Global Knowledge with Project Symmetry Grey-Box is a social innovation startup dedicated to addressing global challenges through technology. Our latest initiative, Project Symmetry, leverages Artificial Intelligence (AI) to accelerate Wikipedia's translation efforts in less-represented languages (those with fewer than 1 million articles). By building a semantic understanding of similar articles in various Wikipedia languages, Project Symmetry provides relevant translations for missing information, promoting inclusivity and equal access to knowledge. [More about the company here: https://www.grey-box.ca/] The Challenge: Language barriers in knowledge dissemination can significantly hinder global understanding and education. This problem is especially prevalent in less-represented languages on Wikipedia, where the amount of available content is limited. Project Symmetry aims to solve this by developing a comprehensive, AI-driven translation tool that ensures accurate and contextually relevant translations, bridging gaps in knowledge accessibility. Grey-Box is a nonprofit structured under the principle of mobile-first and remote-first. Diversity is not only encouraged, it is an integral part of the strength of the team and its projects.

Students will be responsible for delivering the following artifacts by the end of the two-semester sequence: -Improved UI/UX Design: Continue implementing the UI from the in-place Figma design, leveraging React and interfacing with the Symmetry middleware API. -Standalone Application Enhancements: More work on building the layer bundling the React UI and the Python backend into a standalone executable (target platforms include Windows and Mac). -Middleware Integration: More work on the middleware wiring between the UI and the comparison processing. -Comparison Module: Develop a comparison module leveraging AI, similarity metrics, and other tools to analyze provided source and translated content. -Translation Capabilities: Add offline and online translation functionalities focusing on under-represented languages. -AI Model Integration: Identify, select, and integrate AI models, possibly training new ones if it makes sense. -Documentation: Comprehensive documentation of the code, APIs, and usage guidelines. Research Findings: A report detailing the research findings, model evaluations, and future recommendations.