Constructing effective representation of lesions is essential for disease classification and localization in medical image analysis. Prototype-based models address this by leveraging visual prototypes to capture representative lesion patterns, yet effectively handling the complexity of diverse lesion characteristics remains a critical challenge, as they typically rely on single-level, fixed-size prototypes and suffer from prototype redundancy. In this paper, we present HierProtoPNet, a new prototype-based framework designed to handle the complexity of lesions in medical images. HierProtoPNet leverages hierarchical visual prototypes across different semantic feature granularities to effectively capture diverse lesion patterns. To prevent redundancy and increase utility of the prototypes, we devise a novel prototype mining paradigm to progressively discover semantically distinct prototypes, offering multi-level complementary analysis of lesions. Also, we introduce a dynamic knowledge distillation strategy that allows transferring essential classification information across hierarchical levels, thereby improving generalisation performance. Comprehensive experiments show that HierProtoPNet achieves state-of-the-art classification performances in three benchmarks: binary breast cancer screening, multi-class retinal disease diagnosis, and multi-label chest X-ray classification. Quantitative assessments also illustrate HierProtoPNet's significant advantages in weakly-supervised disease localisation and segmentation.

In this paper, we develop a method for extracting information from Large Language Models (LLMs) with associated confidence estimates. We propose that effective confidence models may be designed using a large number of uncertainty measures (i.e., variables that are only weakly predictive of - but positively correlated with - information correctness) as inputs. We trained a confidence model that uses 20 handcrafted uncertainty measures to predict GPT-4’s ability to reproduce species occurrence data from iDigBio and found that, if we only consider occurrence claims that are placed in the top 30% of confidence estimates, we can increase prediction accuracy from 57% to 88% for species absence predictions and from 77% to 86% for species presence predictions. Using the same confidence model, we used GPT-4 to extract new data that extrapolates beyond the occurrence records in iDigBio and used the results to visualize geographic distributions for four individual species. More generally, this represents a novel use case for LLMs in generating credible pseudo data for applications in which high-quality curated data are unavailable or inaccessible.

The deployment of automated deep-learning classifiers in clinical practice has the potential to streamline the diagnosis process and improve the diagnosis accuracy, but the acceptance of those classifiers relies on both their accuracy and interpretability. In general, accurate deep-learning classifiers provide little model interpretability, while interpretable models do not have competitive classification accuracy. In this paper, we introduce a new deep-learning diagnosis framework, called InterNRL, that is designed to be highly accurate and interpretable. InterNRL consists of a student-teacher framework, where the student model is an interpretable prototype-based classifier (ProtoPNet) and the teacher is an accurate global image classifier (GlobalNet). The two classifiers are mutually optimised with a novel reciprocal learning paradigm in which the student ProtoPNet learns from optimal pseudo labels produced by the teacher GlobalNet, while GlobalNet learns from ProtoPNet's classification performance and pseudo labels. This reciprocal learning paradigm enables InterNRL to be flexibly optimised under both fully- and semi-supervised learning scenarios, reaching state-of-the-art classification performance in both scenarios for the tasks of breast cancer and retinal disease diagnosis. Moreover, relying on weakly-labelled training images, InterNRL also achieves superior breast cancer localisation and brain tumour segmentation results than other competing methods.