Diabetic retinopathy (DR) is a major cause of visual impairment, and effective treatment options depend heavily on timely and accurate diagnosis. Deep learning models have demonstrated great success identifying DR from retinal images. However, relying only on predictions made by models, without any indication of model confidence, creates uncertainty and poses significant risk in clinical settings. This paper investigates an alternative in uncertainty-aware deep learning models, including a rejection mechanism to reject low-confidence predictions, contextualized by deferred decision-making in clinical practice. The results show there is a trade-off between prediction coverage and coverage reliability. The Variational Bayesian model adopted a more conservative strategy when predicting DR, subsequently rejecting the uncertain predictions. The model is evaluated by means of important performance metrics such as Accuracy on accepted predictions, the proportion of accepted cases (coverage), the rejection-ratio, and Expected Calibration Error (ECE). The findings also demonstrate a clear trade-off between accuracy and caution, establishing that the use of uncertainty estimation and selective rejection improves the model's reliability in safety-critical diagnostic use cases.

Predictive models are being increasingly used across a wide range of domains, including safety-critical applications such as medical diagnosis and criminal justice. Reliable uncertainty estimation is a crucial task in such settings. Tabular Prior-data Fitted Network (TabPFN) is a recently proposed machine learning foundation model for tabular dataset, which uses a generative transformer architecture. Variational Bayesian Last Layers (VBLL) is a state-of-the-art lightweight variational formulation that effectively improves uncertainty estimation with minimal computational overhead. In this work we aim to evaluate the performance of VBLL integrated with the recently proposed TabPFN in uncertainty calibration. Our experiments, conducted on three benchmark medical tabular datasets, compare the performance of the original TabPFN and the VBLL-integrated version. Contrary to expectations, we observed that original TabPFN consistently outperforms VBLL integrated TabPFN in uncertainty calibration across all datasets.

Gaussian Processes (GPs) are widely seen as the state-of-the-art surrogate models for Bayesian optimization (BO) due to their ability to model uncertainty and their performance on tasks where correlations are easily captured (such as those defined by Euclidean metrics) and their ability to be efficiently updated online. However, the performance of GPs depends on the choice of kernel, and kernel selection for complex correlation structures is often difficult or must be made bespoke. While Bayesian neural networks (BNNs) are a promising direction for higher capacity surrogate models, they have so far seen limited use due to poor performance on some problem types. In this paper, we propose an approach which shows competitive performance on many problem types, including some that BNNs typically struggle with. We build on variational Bayesian last layers (VBLLs), and connect training of these models to exact conditioning in GPs. We exploit this connection to develop an efficient online training algorithm that interleaves conditioning and optimization. Our findings suggest that VBLL networks significantly outperform GPs and other BNN architectures on tasks with complex input correlations, and match the performance of well-tuned GPs on established benchmark tasks.