RATNet applies analogical reasoning via a cyclic pre-training strategy to outperform prior foundation models in GI endoscopy diagnosis across diagnosis, few-shot, zero-shot, robustness, adaptation, and federated scenarios.
CA: a cancer journal for clinicians71(3), 209–249 (2021) 17
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Analogical Reasoning as a Doctor: A Foundation Model for Gastrointestinal Endoscopy Diagnosis
RATNet applies analogical reasoning via a cyclic pre-training strategy to outperform prior foundation models in GI endoscopy diagnosis across diagnosis, few-shot, zero-shot, robustness, adaptation, and federated scenarios.