Tyee: A Unified, Modular, and Fully-Integrated Configurable Toolkit for Intelligent Physiological Health Care

TL;DR

Tyee tackles the fragmentation and reproducibility challenges in physiological signal analysis by delivering a unified, modular toolkit with a configurable data interface, extensible components, and end-to-end workflow automation. The approach centers on a four-layer architecture (Entity, Task, Training, Configuration) and YAML-based configuration to enable rapid prototyping, fair benchmarking, and scalable experimentation across 12 signal modalities and 11 healthcare tasks. Empirical results show Tyee matching or surpassing baselines on 12 of 13 datasets, highlighting strong generalizability in unimodal and multimodal settings. As an open-source platform with broad dataset support and easy deployment, Tyee is positioned to accelerate research, collaboration, and deployment of intelligent physiological health solutions.

Abstract

Deep learning has shown great promise in physiological signal analysis, yet its progress is hindered by heterogeneous data formats, inconsistent preprocessing strategies, fragmented model pipelines, and non-reproducible experimental setups. To address these limitations, we present Tyee, a unified, modular, and fully-integrated configurable toolkit designed for intelligent physiological healthcare. Tyee introduces three key innovations: (1) a unified data interface and configurable preprocessing pipeline for 12 kinds of signal modalities; (2) a modular and extensible architecture enabling flexible integration and rapid prototyping across tasks; and (3) end-to-end workflow configuration, promoting reproducible and scalable experimentation. Tyee demonstrates consistent practical effectiveness and generalizability, outperforming or matching baselines across all evaluated tasks (with state-of-the-art results on 12 of 13 datasets). The Tyee toolkit is released at https://github.com/SmileHnu/Tyee and actively maintained.

Figures (2)

• Figure 1:
• Figure 2: Radar plots comparing Tyee (red) vs. baselines (blue) in (a) unimodal and (b) multimodal healthcare tasks. Tyee matches or outperforms the baselines on 12 out of 13 datasets, demonstrating reliable and consistent benchmark reproduction.