[Edaily Reporter Song Young-doo] Wearable AI diagnostic monitoring company #Ciers announced on the 11th that its AI technology for detecting abnormal vital signs in hospitalized patients has been accepted by "ICLR 2026" (International Conference on Learning Representations), one of the world’s most prestigious AI academic conferences.

The research was presented under the title "Enhancing Sparse Event Detection in Healthcare Time-Series via Adaptive Gate of Context-detail Interaction."

ICLR serves as an academic platform where researchers from major global research institutions and big tech companies compete to showcase the latest AI technological achievements. With this acceptance, Cears is recognized internationally for its competitiveness in core medical AI technology.

In this study, the C-EARS research team proposed an AI framework that precisely detects the types and occurrence intervals of abnormal events in real-time biosignal data, including electrocardiograms (ECGs).

Existing AI models had limitations in accurately identifying so-called "sparse events"—events with low clinical frequency and unclear boundaries. Ciers enhanced both detection performance and interpretability by combining "Adaptive Gating" technology with a new AI architecture that simultaneously learns global contextual information (capturing the overall signal flow) and detailed waveform features.

Research results demonstrated that this technology outperformed existing models across various time-series datasets, including arrhythmia detection, emotion recognition, and human activity monitoring. It is evaluated as having enhanced its applicability in clinical settings by enabling the more accurate detection of clinically significant abnormal signals, going beyond simply determining the presence of an anomaly.

These research findings directly contribute to the enhancement of the alarm system in thynC™, Cears’ AI-based inpatient monitoring platform. thynC™ is a platform that uses AI analysis to detect changes in a patient’s condition in real time and provides warning alerts to medical staff when abnormal signs occur. This technology, which detects rare abnormal signals with greater precision, is expected to contribute to improving early detection accuracy and enhancing AI-based clinical decision support capabilities.

Ciers currently holds over 120 domestic and international patents, as well as numerous clinical studies and academic papers, and is advancing its AI prediction and analysis technologies based on the large-scale medical data accumulated through thynC™ and Mobicare.

Song Hee-seok, Executive Vice President and Chief Technology Officer (CTO) of Cears, stated, “This research is based on the data assets and clinical experience we have accumulated over the years and can be utilized for AI-based clinical decision-making in inpatient wards.” He added, “We will continue to strengthen the competitiveness of our core AI technologies and expand our AI services based on time-series data.”