Prototypical Calibration

Prototypical calibration is a technique used to improve the accuracy and robustness of machine learning models, particularly in scenarios with limited data or noisy inputs. Current research focuses on applying this approach across diverse fields, including medical image segmentation, nuclear power plant monitoring, and few-shot learning for language models, often employing neural networks (like convolutional neural networks) and leveraging techniques like contrastive learning and Gaussian mixture models to refine model predictions. This work aims to enhance model generalization and reduce uncertainty, leading to more reliable predictions in various applications where data scarcity or heterogeneity poses challenges. The resulting improvements in model performance have significant implications for both scientific understanding and practical applications across multiple domains.

Papers

June 27, 2024

Data-Driven Prediction and Uncertainty Quantification of PWR Crud-Induced Power Shift Using Convolutional Neural Networks
Aidan Furlong, Farah Alsafadi, Scott Palmtag, Andrew Godfrey, Xu Wu
Convolutional Neural Network Uncertainty Quantification Data Driven Model Prediction Nuclear Reactor Prototypical Calibration

January 25, 2024

ProCNS: Progressive Prototype Calibration and Noise Suppression for Weakly-Supervised Medical Image Segmentation
Y. Liu, L. Lin, K. K. Y. Wong, X. Tang
Pseudo Label Soft Label Sparse Annotation Sparse Learning Noise Suppression Prototypical Calibration

December 8, 2023

Few-Shot Class-Incremental Learning via Training-Free Prototype Calibration
Qi-Wei Wang, Da-Wei Zhou, Yi-Kai Zhang, De-Chuan Zhan, Han-Jia Ye
Class Incremental Learning Unseen Class General Class Prototypical Calibration

August 7, 2023

Cross-Silo Prototypical Calibration for Federated Learning with Non-IID Data
Zhuang Qi, Lei Meng, Zitan Chen, Han Hu, Hui Lin, Xiangxu Meng
Raw Data Prototypical Calibration

May 20, 2022

Prototypical Calibration for Few-shot Learning of Language Models
Zhixiong Han, Yaru Hao, Li Dong, Yutao Sun, Furu Wei
Language Model Context Learning Shot Classification Different Cluster Robust Decision Prototypical Calibration

Prototypical Calibration

Papers

Data-Driven Prediction and Uncertainty Quantification of PWR Crud-Induced Power Shift Using Convolutional Neural Networks

ProCNS: Progressive Prototype Calibration and Noise Suppression for Weakly-Supervised Medical Image Segmentation

Few-Shot Class-Incremental Learning via Training-Free Prototype Calibration

Cross-Silo Prototypical Calibration for Federated Learning with Non-IID Data

Prototypical Calibration for Few-shot Learning of Language Models