InFL-UX: A Toolkit for Web-Based Interactive Federated Learning
Tim Maurer, Abdulrahman Mohamed Selim, Hasan Md Tusfiqur Alam, Matthias Eiletz, Michael Barz, Daniel Sonntag
TL;DR
The paper tackles privacy constraints in ML by integrating federated learning with interactive user input in a browser. It introduces InFL-UX, a browser-based PoC toolkit that enables in-browser, asynchronous training using $FedAsync$, ONNX Runtime, WebAssembly, and WebGPU, with users uploading datasets and refining labels. Key contributions include a two-server architecture, administrator-configurable training, Docker deployment, and a demonstration on image classification with $MobileNetV2$. Limitations include incomplete on-web training in ONNX Runtime (e.g., fixed learning rate $0.001$ and limited losses) and browser/storage constraints, with future work on broader CV tasks, explainable AI, and multi-task deployments. Overall, the work advances practical, user-centric FL in privacy-sensitive domains by enabling collaborative, in-browser training across devices.
Abstract
This paper presents InFL-UX, an interactive, proof-of-concept browser-based Federated Learning (FL) toolkit designed to integrate user contributions seamlessly into the machine learning (ML) workflow. InFL-UX enables users across multiple devices to upload datasets, define classes, and collaboratively train classification models directly in the browser using modern web technologies. Unlike traditional FL toolkits, which often focus on backend simulations, InFL-UX provides a simple user interface for researchers to explore how users interact with and contribute to FL systems in real-world, interactive settings. By prioritising usability and decentralised model training, InFL-UX bridges the gap between FL and Interactive Machine Learning (IML), empowering non-technical users to actively participate in ML classification tasks.