Edge AI for Early Detection of Chronic Diseases and the Spread of Infectious Diseases: Opportunities, Challenges, and Future Directions

Future Internet, Год журнала: 2023, Номер 15(11), С. 370 - 370

Опубликована: Ноя. 18, 2023

Edge AI, an interdisciplinary technology that enables distributed intelligence with edge devices, is quickly becoming a critical component in early health prediction. AI encompasses data analytics and artificial (AI) using machine learning, deep federated learning models deployed executed at the of network, far from centralized centers. careful analysis large datasets derived multiple sources, including electronic records, wearable demographic information, making it possible to identify intricate patterns predict person’s future health. Federated novel approach further enhances this prediction by enabling collaborative training on devices while maintaining privacy. Using computing, can be processed analyzed locally, reducing latency instant decision making. This article reviews role highlights its potential improve public Topics covered include use algorithms for detection chronic diseases such as diabetes cancer computing detect spread infectious diseases. In addition discussing challenges limitations prediction, emphasizes research directions address these concerns integration existing healthcare systems explore full technologies improving

Язык: Английский

---

Learning under label noise through few-shot human-in-the-loop refinement

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Фев. 4, 2025

Wearable technologies enable continuous monitoring of various health metrics, such as physical activity, heart rate, sleep, and stress levels. A key challenge with wearable data is obtaining quality labels. Unlike modalities like video where the videos themselves can be effectively used to label objects or events, do not contain obvious cues about manifestation users usually require rich metadata. As a result, noise become an increasingly thorny issue when labeling data. In this paper, we propose novel solution address noisy learning, entitled Few-Shot Human-in-the-Loop Refinement (FHLR). Our method initially learns seed model using weak Next, it fine-tunes handful expert corrections. Finally, achieves better generalizability robustness by merging fine-tuned models via weighted parameter averaging. We evaluate our approach on four challenging tasks datasets, compare against eight competitive baselines designed deal show that FHLR significantly performance learning from labels state-of-the-art large margin, up $$19\%$$ accuracy improvement under symmetric asymmetric noise. Notably, find particularly robust increased noise, unlike prior works suffer severe degradation. work only generalization in high-stakes sensing benchmarks but also sheds light how affects commonly-used models.

Язык: Английский

---

Digital health technology in clinical trials

npj Digital Medicine, Год журнала: 2023, Номер 6(1)

Опубликована: Май 18, 2023

Digital health technologies (DHTs) have brought several significant improvements to clinical trials, enabling real-world data collection outside of the traditional context and more patient-centered approaches. DHTs, such as wearables, allow unique personal at home over a long period. But DHTs also bring challenges, digital endpoint harmonization disadvantaging populations already experiencing divide. A recent study explored growth trends implications established novel in neurology trials past decade. Here, we discuss benefits future challenges DHT usage trials.

Язык: Английский

---

Occupant-centered indoor environmental quality management: Physiological response measuring methods

Building and Environment, Год журнала: 2023, Номер 243, С. 110661 - 110661

Опубликована: Июль 24, 2023

Язык: Английский

---

Cyclical underreporting of seizures in patient‐based seizure documentation

Annals of Clinical and Translational Neurology, Год журнала: 2023, Номер 10(10), С. 1863 - 1872

Опубликована: Авг. 23, 2023

Circadian and multidien cycles of seizure occurrence are increasingly discussed as to their biological underpinnings in the context forecasting. This study analyzes if patient reported seizures provide valid data on such cyclical occurrence.We retrospectively studied circadian derived from patient-based reporting reflect objective documentation 2003 patients undergoing in-patient video-EEG monitoring.Only 24.1% more than 29000 documented were accompanied by notifications. There was underreporting with a maximum during nighttime, leading significant deviations distribution seizures. Significant found for focal epilepsies originating both, frontal temporal lobes, different types (in particular, unaware bilateral tonic-clonic seizures).Patient diaries may bias rather true distributions. Cyclical reports alone lead suboptimal treatment schemes, an underestimation seizure-associated risks, pose problems finding strongly supports use measures monitor distributions studies decisions based thereon.

Язык: Английский

---

Evaluation of Data Processing and Artifact Removal Approaches Used for Physiological Signals Captured Using Wearable Sensing Devices during Construction Tasks

Journal of Construction Engineering and Management, Год журнала: 2023, Номер 150(1)

Опубликована: Окт. 25, 2023

Wearable sensing devices (WSDs) have enormous promise for monitoring construction worker safety. They can track workers and send safety-related information in real time, allowing more effective preventative decision making. WSDs are particularly useful on sites since they workers' health, safety, activity levels, among other metrics that could help optimize their daily tasks. may also assist recognizing health-related safety risks (such as physical fatigue) taking appropriate action to mitigate them. The data produced by these WSDs, however, is highly noisy contaminated with artifacts been introduced the surroundings, experimental apparatus, or subject's physiological state. These very strong frequently found during field experiments. So, when there a lot of artifacts, signal quality drops. Recently, removal has greatly enhanced developments processing, which vastly performance. Thus, proposed review aimed provide an in-depth analysis approaches currently used analyze remove from signals obtained via construction-related First, this study provides overview likely be recorded monitor health Second, identifies most prevalent detrimental effect utility signals. Third, comprehensive existing artifact-removal were presented. Fourth, each identified artifact detection approach was analyzed its strengths weaknesses. Finally, conclusion, few suggestions future research improving captured using approaches.

Язык: Английский

---

Photoplethysmography signal quality assessment using attractor reconstruction analysis

Biomedical Signal Processing and Control, Год журнала: 2023, Номер 86, С. 105142 - 105142

Опубликована: Июнь 19, 2023

Язык: Английский

---

Home recording of 3‐Hz spike–wave discharges in adults with absence epilepsy using the wearable Sensor Dot

Epilepsia, Год журнала: 2023, Номер 65(2), С. 378 - 388

Опубликована: Дек. 1, 2023

Home monitoring of 3-Hz spike-wave discharges (SWDs) in patients with refractory absence epilepsy could improve clinical care by replacing the inaccurate seizure diary objective counts. We investigated use and performance Sensor Dot (Byteflies) wearable persons their home environment.

Язык: Английский

---

A real-world dataset of group emotion experiences based on physiological data

Scientific Data, Год журнала: 2024, Номер 11(1)

Опубликована: Янв. 23, 2024

Abstract Affective computing has experienced substantial advancements in recognizing emotions through image and facial expression analysis. However, the incorporation of physiological data remains constrained. Emotion recognition with shows promising results controlled experiments but lacks generalization to real-world settings. To address this, we present G-REx, a dataset for affective computing. We collected (photoplethysmography electrodermal activity) using wrist-worn device during long-duration movie sessions. annotations were retrospectively performed on segments elevated responses. The includes over 31 sessions, totaling 380 h+ from 190+ subjects. group setting, which can give further context emotion systems. Our setup aims be easily replicable any real-life scenario, facilitating collection large datasets novel

Язык: Английский

---

The Importance of Data Quality Control in Using Fitbit Device Data From the Research Program

JMIR mhealth and uhealth, Год журнала: 2023, Номер 11, С. e45103 - e45103

Опубликована: Сен. 8, 2023

Abstract Wearable digital health technologies (DHTs) have become increasingly popular in recent years, enabling more capabilities to assess behaviors and physiology free-living conditions. The All of Us Research Program (AoURP), a National Institutes Health initiative that collects health-related information from participants the United States, has expanded its data collection include DHT Fitbit devices. This offers researchers an unprecedented opportunity examine large cohort alongside biospecimens electronic records. However, there are existing challenges sources error need be considered before using device AoURP. In this viewpoint, we reliability potential associated with available through AoURP Researcher Workbench outline actionable strategies mitigate missingness noise. We begin by discussing noise, including (1) inherent measurement inaccuracies, (2) skin tone–related challenges, (3) movement motion artifacts, proceed discuss data. then methods such noise end considering how future enhancements AoURP’s inclusion new types would impact usability Although considerations suggested literature tailored toward AoURP, recommendations broadly applicable wearable DHTs

Язык: Английский

---