Accepted Papers to ISWC26

📢 We at #TECO are proud to announce that three of our papers have been accepted to #ISWC26 at #UbiCompISWC26 in #Shanghai! This year, our contributions span personalized #HumanActivityRecognition, robust evaluation for wearable health sensing, and earbud-based cardiac monitoring.

---

🏃 Personalized activity recognition with just seconds of calibration
“BayaHAR: Lightweight Bayesian Few-Shot User Adaptation for On-Device Personalized Human Activity Recognition” – Maximilian Burzer et al.

With #BayaHAR, we introduce a lightweight, gradient-free approach for personalizing pretrained HAR models directly on wearable devices. Bayesian prototype updates enable adaptation from very little user-specific data: with only three seconds of calibration per activity, supervised adaptation improved macro-F1 by 2.76 to 33.44 percentage points across four datasets, while requiring no backpropagation on-device.

📄 Preprint: https://lnkd.in/dS7Sa89j

---

🩸 When promising glucose-estimation results do not generalize
“Reassessing the Feasibility of PPG-Based Non-Invasive Blood Glucose Level Estimation” – Supraja Ramesh et al.

We systematically reassess five representative methods for estimating blood glucose from #PPG using a reproducible evaluation pipeline. While models appear competitive with random train-test splits, their performance largely collapses when evaluated on unseen participants. At the same time, commonly used clinical error-grid metrics still report more than 90% acceptable predictions – even for a simple baseline – highlighting why careful evaluation is essential before assessing real-world utility.

📄 Preprint: https://lnkd.in/dQg5jvjY

---

❤️ Touch your earbud to capture an ECG
“Finger-to-Ear ECG: Systematic Evaluation of an Earbud-Based Cardiac Monitoring Approach” – Philipp Lepold et al.

#FingerToEarECG combines an in-ear electrode with a finger-contact electrode on the earbud housing for opportunistic cardiac sensing. In a study with 30 participants, left-finger configurations achieved R-peak detection F1-scores above 99% while preserving important ECG morphology, demonstrating the potential of a simple earbud touch interaction for high-quality cardiac monitoring.

📄 Preprint: https://lnkd.in/dJGYuK8w

---

#ISWC #WearableComputing #WearableTech #HumanActivityRecognition #PersonalizedAI #PPG #Glucose #DigitalHealth #Earables #CardiacMonitoring #MachineLearning #HumanComputerInteraction #TECO