Unicode, Inc.

Mentors: Student: Preet Sojitra

The Unicode Consortium maintains extensive locale datasets critical for internationalization tasks such as date/time formatting, text segmentation, measurement conversions, and collation. Manual verification of these datasets is challenging due to their vast scale. My project aims to address this by leveraging large language models (LLMs) to automate and enhance the validation of locale data. The approach involves designing targeted queries for an LLM, systematically comparing the generated outputs with existing CLDR/ICU datasets, and identifying discrepancies. These results will then be processed to isolate significant issues, and automated tools will be developed to flag potentially erroneous data for translators. Deliverables: 1.A robust pipeline for querying LLMs and generating locale data predictions. 2.Comprehensive comparative analysis between LLM predictions and CLDR/ICU datasets. 3.An automated issue detection and reporting mechanism integrated with existing Unicode systems. 4.Thorough documentation and guidelines for future dataset validation workflows.

Mentors: Student: Hussain A.

The Unicode Inflection project currently provides its code in source form, which can be a challenge for users and developers who want a simple, ready-to-use solution. The goal of this project is to create automated, cross-platform packaging for the library, making it easier for users to install and use across Linux, macOS, and Windows. I will achieve this by utilizing CMake to streamline the build process, creating CMake-based scripts that can automatically generate compressed source and binary distribution packages for each platform. The final deliverables will include a set of tools to package the library and data files, along with clear documentation on how to build and distribute the packages. This work will improve accessibility for new users, save time for developers, and enhance the overall usability of the Unicode Inflection project.

Mentors: Student: Popoola Precious

This proposal focuses on developing two essential tools using Flutter: the Unicode Character Visualizer App and the Multilingual Text Rendering Tester App. Problem Statement Many developers, designers, and linguists lack accessible tools to browse Unicode characters, view detailed metadata, and compare how multilingual text is rendered across devices and fonts. Text rendering inconsistencies often arise across platforms, impacting accessibility and readability. Solution Unicode Character Visualizer App: Provide a comprehensive and interactive platform to search and visualize Unicode characters. Display essential metadata such as codepoints, descriptions, and language support. Offer font rendering previews for accurate character representation. Multilingual Text Rendering Tester App: Allow users to input multilingual text and visualize rendering differences across various fonts and devices. Benchmark rendering performance and identify inconsistencies. Deliverables Fully functional Unicode Character Visualizer App with search, character details, and font rendering features. Multilingual Text Rendering Tester App with input comparison, font analysis, and reporting functionalities. Comprehensive documentation, including user guides and API integration manuals. Well-tested codebase with unit, widget, and integration tests. This project will contribute to Unicode, Inc. by providing essential tools for ensuring Unicode character accuracy, supporting linguistic diversity, and improving cross-platform text rendering. I am excited about the opportunity to collaborate and contribute to the Unicode ecosystem.

Mentors: Student: Shenghong Liu

With the emergence of Internet of Things devices, lightweight models have become increasingly vital. This project focuses on building computationally efficient text segmentation models, enabling them to run effectively on devices with limited resources like memory, processing power, and battery life. I aim to improve the current model and expand support to 2-3 additional languages, potentially benefitting hundreds of millions more people. This will be achieved by restructuring the model using TensorFlow's Functional API, which facilitates the implementation of Conditional Random Fields. Additionally, I will be creating an intuitive model evaluation script to offer a customizable experience for selecting the most suitable model.

Mentors: Student: Om Doiphode

Attackers can exploit homoglyphs, or visually similar characters, to perform social engineering attacks or evade spam and plagiarism detectors. Unicode contains over 200 writing systems (scripts). Many of these scripts, such as Latin (in western Europe) and Cyrillic (in eastern Europe), contain characters that are look-alikes. This project aims to develop an AI tool to detect confusable characters and ultimately improve Unicode security mechanisms automatically.