Evidence
Structured evidence behind NeuTigers' platform and programs — clinical and real-world studies, foundational research publications, and media recognition, each with an explicit status.
How to read this page
Every record below states what kind of evidence it is (peer-reviewed, preprint, study or media), who was involved, and the status of the related program. Evidence of research or study activity is not evidence of regulatory clearance or diagnostic status; see the disclaimer in the site footer.
Clinical & real-world evidence
SweetDeep™ — real-time type 2 diabetes detection with a smartwatch
- Year: 2026 · Type: Preprint (arXiv) · Collaborator: AP-HP Avicennes–Jean Verdier–René Muret · Status: Prospective study
- NeuTigers published SweetDeep on arXiv, describing real-time type 2 diabetes detection with a smartwatch — no blood draw required — using data collected in real-world conditions in a decentralized clinical trial with AP-HP.
SweetDeep — clinical validation versus questionnaire-based screening (FINDRISC)
- Year: 2026 · Type: Clinical validation (4 studies, 3 continents: US/EU/MENA) · Collaborators: AP-HP (Paris), CHSF, Hackensack Meridian Health, Jersey Shore University Medical Center, CIC Clinical Research Center & Burjeel Hospitals Network, Princeton University · Status: Prospective studies, completed and ongoing
- Across 4 prospective clinical studies on 3 continents, SweetDeep outperforms standard questionnaire-based risk screening (FINDRISC) with a stronger balance of sensitivity and specificity, enabling earlier identification of at-risk individuals and more efficient referral for confirmatory laboratory testing.
MFDeep — fetal macrosomia risk linked to gestational diabetes
- Year: 2026 · Type: Study / program record · Collaborators: Genopole, CHSF; supported by France 2030 · Status: Active clinical development
- MFDeep predicts fetal macrosomia risk linked to gestational diabetes with real-time, on-device inference delivered to medical teams; the program is featured on the French government digital-health innovation portal gnius.esante.gouv.fr.
MFDeep — France 2030 non-dilutive R&D grant
- Year: 2026 · Type: Funding record · Collaborator: Genopole · Status: Awarded
- MFDeep was awarded a $300K non-dilutive R&D grant via the France 2030 program with Genopole, supporting the program's clinical development in women's health.
Sickle cell disease — wearables study with AP-HP
- Year: 2021 · Type: Study · Collaborator: AP-HP (Hôpitaux de Paris) · Status: Clinical research
- NeuTigers and AP-HP announced a study investigating the use of AI and everyday wearables, such as the Withings ScanWatch, to help prevent hospitalizations of patients with sickle cell anemia by anticipating vaso-occlusive crises. Announcement
CovidDeep — COVID-19 detection from smartwatch sensors
- Year: 2021 · Type: Peer-reviewed (IEEE Transactions on Consumer Electronics) · Collaborators: Princeton University research team; European clinical sites · Status: Published research (historical)
- CovidDeep detected SARS-CoV-2 infection from commercially available wearable sensors and questionnaire data, built on a controlled prospective study of about 2,000 participants in Europe; the reported results are those of the publication and do not represent a cleared diagnostic. About the paper
MHDeep — mental health disorder detection (proof of concept)
- Year: 2021 · Type: Research publication · Collaborator: Princeton University research lineage · Status: Research
- MHDeep demonstrated a proof of concept for detecting three mental health disorders (schizoaffective, major depressive and bipolar disorders) from smartwatch and smartphone sensor data, with reported accuracies of 82.4-90.4% in the published research setting.
Foundational research
Selected research publications underpinning the StarDeep platform, from the Princeton-origin research lineage. PDFs for several of these are available under /assets/docs/ where applicable.
- Grow-and-Prune — Research publication — training method producing compact, efficient deep neural networks for edge inference.
- NeST — Research publication — neural network synthesis combining growth and pruning.
- ChamNet — Research publication — platform-aware neural architecture adaptation.
- SCANN — Research publication — synthesis of compact accurate neural networks.
- LSTM-based modeling — Research publication — recurrent modeling of physiological time-series signals.
- TUTOR — Research publication — combining synthetic and real data to improve model training.
- CTRL — Research publication — detecting labeling errors to improve training-data quality.
- SCouT — Research publication — toward time- and subject-efficient clinical studies from spatiotemporal health data.
- DiabDeep — Research publication — the earlier wearables-based diabetes detection research lineage that preceded the current SweetDeep program.
- SHARKS / GRAVITAS — Research publications — machine-learning-based security analysis for connected and cyber-physical systems (defensibility layer, not a health program).
Patents & IP portfolio
- Type: Intellectual property record · Status: Granted and filed, 2018–2027
- NeuTigers' platform is protected by granted and filed patents covering its core edge-AI methods (including Grow-and-Prune and SCANN) and disease applications (including DiabDeep and MHDeep), within a broader 2018–2027 IP portfolio. See the Platform page for the patent list.
Media & recognition
Fast Company — Next Big Things In Tech 2021
- Year: 2021 · Type: Media recognition · Status: Published
- NeuTigers, a Princeton University spinout building on-device Edge-AI for real-time disease detection, was named in Fast Company's Next Big Things In Tech 2021. Coverage
CovidDeep — most popular paper in IEEE Transactions on Consumer Electronics
- Year: 2021 · Type: Media / recognition · Status: Published
- The CovidDeep paper became the most popular paper in IEEE Transactions on Consumer Electronics. Announcement
gnius.esante.gouv.fr — MFDeep feature
- Year: 2026 · Type: Media (French government e-health innovation portal) · Status: Published
- MFDeep and the Genopole/CHSF partnership are featured on gnius.esante.gouv.fr, the French government's digital-health innovation portal.
For partnership announcements (Samsung 2023, TII 2023, EuroBiomed 2022), see Company news and the partner page.