On Performance and Calibration of Natural Gradient Langevin Dynamics
Producing deep neural network (DNN) models with calibrated confidence is essential for applications in many fields, such as medical image analysis, natural language processing, and robotics. Modern neural networks have been reported to be poorly calibrated compared with those from a decade ago. The...
Главные авторы: | Hanif Amal Robbani, Alhadi Bustamam, Risman Adnan, Shandar Ahmad |
---|---|
Формат: | Статья |
Язык: | English |
Опубликовано: |
IEEE
2023-01-01
|
Серии: | IEEE Access |
Предметы: | |
Online-ссылка: | https://ieeexplore.ieee.org/document/10131934/ |
Схожие документы
-
Fast Sampling of Score-Based Models With Cyclical Diffusion Sampling
по: Karimul Makhtidi, и др.
Опубликовано: (2024-01-01) -
On a generalization of fractional Langevin equation with boundary conditions
по: Zheng Kou, и др.
Опубликовано: (2022-01-01) -
Dynamical Sampling with Langevin Normalization Flows
по: Minghao Gu, и др.
Опубликовано: (2019-11-01) -
Calibration with confidence: a principled method for panel assessment
по: R. S. MacKay, и др.
Опубликовано: (2017-01-01) -
Lévy-walk-like Langevin dynamics
по: Xudong Wang, и др.
Опубликовано: (2019-01-01)