Retract
Screen your paper’s retraction risk

How to use Retract

1. Paste only the abstract into the text box (no title, no keywords, no methods/results/discussion, no full text).
2. Choose a threshold (slider). Higher threshold = stricter flagging.
3. Click Check.
4. Get result: the score is the probability (%) that the abstract shows patterns similar to cardiovascular-medicine original research papers retracted within the last ~5 years. (Model is trained on cardiovascular papers; see “About” for details.)

Ads are shown to free users. You can remove ads for this app by spending 10 credits (Account → Remove ads).

About Retract

Retract is a lightweight screening tool that estimates how strongly an abstract resembles cardiovascular-medicine papers retracted within the last ~5 years. It is powered by a fine-tuned BERT text-classification model trained on 520 retracted research articles (2021–2026) from the Retraction Watch Database, using cardiovascular/vascular/interventional/venous/dialysis/stroke/coronary-related keywords. The control set consists of top-level publications matched by similar keywords from PubMed. The model is hosted on Hugging Face.

Model training: AutoTrain (Text Classification).
Validation metrics: loss 0.346 · F1 0.862 · precision 0.955 · recall 0.785 · AUC 0.920 · accuracy 0.875

Credits are universal across MLHS apps. 1 credit covers typical single actions (and can be used for ad-free in Retract).

Part of ML in Health Science © .

Version

• 1.0: Initial release (abstract screening + MLHS shared login/credits + ad-free entitlement).

If you report a bug, include a screenshot + browser/device details.