Abstract:Prototypical-part methods, e.g., ProtoPNet, enhance interpretability in image recognition by linking predictions to training prototypes, thereby offering intuitive insights into their decision-making. Existing methods, which rely on a point-based learning of prototypes, typically face two critical issues: 1) the learned prototypes have limited representation power and are not suitable to detect Out-of-Distribution (OoD) inputs, reducing their decision trustworthiness; and 2) the necessary projection of the learned prototypes back into the space of training images causes a drastic degradation in the predictive performance. Furthermore, current prototype learning adopts an aggressive approach that considers only the most active object parts during training, while overlooking sub-salient object regions which still hold crucial classification information. In this paper, we present a new generative paradigm to learn prototype distributions, termed as Mixture of Gaussian-distributed Prototypes (MGProto). The distribution of prototypes from MGProto enables both interpretable image classification and trustworthy recognition of OoD inputs. The optimisation of MGProto naturally projects the learned prototype distributions back into the training image space, thereby addressing the performance degradation caused by prototype projection. Additionally, we develop a novel and effective prototype mining strategy that considers not only the most active but also sub-salient object parts. To promote model compactness, we further propose to prune MGProto by removing prototypes with low importance priors. Experiments on CUB-200-2011, Stanford Cars, Stanford Dogs, and Oxford-IIIT Pets datasets show that MGProto achieves state-of-the-art image recognition and OoD detection performances, while providing encouraging interpretability results.

Propagation Based Prototype Prediction

Real-Time Prediction of TBM Driving Parameters Using Geological and Operation Data<italic/>

Effective Model Integration Algorithm for Improving Link and Sign Prediction in Complex Networks

Prototype-based Semantic Segmentation

Multi-Modal Prototypes for Open-World Semantic Segmentation

Prototype Matching Networks for Large-Scale Multi-label Genomic Sequence Classification

A Hierarchical Prototype-Based Approach for Classification.

Towards Effective Next POI Prediction: Spatial and Semantic Augmentation with Remote Sensing Data

ProtoryNet - Interpretable Text Classification Via Prototype Trajectories

Sparse Prototype Network for Explainable Pedestrian Behavior Prediction

Learning from the Target: Dual Prototype Network for Few Shot Semantic Segmentation

Mixture of Gaussian-distributed Prototypes with Generative Modelling for Interpretable and Trustworthy Image Recognition

D2D-LSTM based Prediction of the D2D Diffusion Path in Mobile Social Networks

MProto: Multi-Prototype Network with Denoised Optimal Transport for Distantly Supervised Named Entity Recognition

Deep Learning for Case-Based Reasoning Through Prototypes: A Neural Network That Explains Its Predictions

POEM: A prototype cross and emphasis network for few-shot semantic segmentation

A Label Propagation Based User Locations Prediction Algorithm in Social Network

Prototype extraction and adaptive OCR

GAProtoNet: A Multi-head Graph Attention-based Prototypical Network for Interpretable Text Classification

PrototypeFormer: Learning to Explore Prototype Relationships for Few-shot Image Classification