Enhancing Sparse Event Detection in Healthcare Time-Series via Adaptive Gate of Context–Detail Interaction

Accurate detection of clinically meaningful events in healthcare time-series data is crucial for reliable downstream analysis and decision support. However, most existing methods struggle to jointly localize event boundaries and classify event types; even detection transformer (DETR)-based approaches show limited performance when confronted with extremely sparse events typical of clinical recordings. To address these challenges, we propose a coarse-to-fine detection framework combining a global context explorer, a local detail inspector, and an adaptive gating module (AGM) that fuses multiple label perspectives. The AGM uses transformed labels-encoding event presence and temporal position-to improve learning on sparse events. This design acts as a switch that selectively activates detailed feature extraction only when an event is likely, thereby reducing noise and improving efficiency in sparse settings. We evaluate our framework on diverse healthcare datasets-including arrhythmia detection, emotion recognition, and human-activity monitoring-and demonstrate substantial performance gains over existing DETR-based models, with particularly strong improvements in sparse event detection. With precise and robust event detection, our framework enables interpretation and actionable insights in real-world clinical applications. * Corresponding author. RELATED WORK SEQUENCE-TO-SEQUENCE MODELS FOR EVENT DETECTION Seq-to-seq models for time-series analysis have evolved rapidly, starting from LSTM (Graves, 2012), TCN (Bai et al., 2018), and Inception Time (Ismail Fawaz et al., 2020), and more recently advancing to Transformer-based architectures (Vaswani et al., 2017; Nie et al., 2022; Liu et al., 2023) and time-series-specific foundation models (Garza et al., 2023; Das et al., 2024) . While these models are effective at capturing global patterns and performing point-wise prediction or regression, they face structural limitations for event boundary detection within segments. Point-wise outputs are inherently insufficient for tasks that require identifying both event onset and offset. Moreover, these models are not well-suited for detecting sparse events, where occurrences are rare and irregular.