Assessing Box Merging Strategies and Uncertainty Estimation Methods in Multimodel Object Detection

2020 ◽  
pp. 3-10
Author(s):  
Felippe Schmoeller Roza ◽  
Maximilian Henne ◽  
Karsten Roscher ◽  
Stephan Günnemann
Keyword(s):  
Object Detection ◽  
Uncertainty Estimation ◽  
Estimation Methods

scholarly journals Evaluating Scalable Uncertainty Estimation Methods for Deep Learning-Based Molecular Property Prediction

2020 ◽  
Vol 60 (6) ◽  
pp. 2697-2717 ◽  
Author(s):  
Gabriele Scalia ◽  
Colin A. Grambow ◽  
Barbara Pernici ◽  
Yi-Pei Li ◽  
William H. Green
Keyword(s):  
Deep Learning ◽  
Uncertainty Estimation ◽  
Estimation Methods ◽  
Molecular Property ◽  
Property Prediction

scholarly journals CertainNet: Sampling-free Uncertainty Estimation for Object Detection

2021 ◽  
pp. 1-1
Author(s):  
Stefano Gasperini ◽  
Jan Haug ◽  
Mohammad-Ali Nikouei Mahani ◽  
Alvaro Marcos-Ramiro ◽  
Nassir Navab ◽  
...  
Keyword(s):  
Object Detection ◽  
Uncertainty Estimation

scholarly journals Exploring Uncertainty in Streamflow Estimates

Eos ◽  
2019 ◽  
Vol 100 ◽  
Author(s):  
Aaron Sidder
Keyword(s):  
Uncertainty Estimation ◽  
Estimation Methods

A review of streamflow uncertainty estimation methods reveals that one method does not fit all situations and provides recommendations for how to improve streamflow estimates.

The death of the Job plot, transparency, open science and online tools, uncertainty estimation methods and other developments in supramolecular chemistry data analysis

2016 ◽  
Vol 52 (87) ◽  
pp. 12792-12805 ◽  
Author(s):  
D. Brynn Hibbert ◽  
Pall Thordarson
Keyword(s):  
Data Analysis ◽  
Open Access ◽  
Best Practice ◽  
Open Science ◽  
Uncertainty Estimation ◽  
Estimation Methods ◽  
Binding Studies ◽  
Feature Article ◽  
Online Tools ◽  
Guest Binding

The failure of the Job plot, best-practice in uncertainty estimation in host–guest binding studies and an open access webportal for data analysis are reviewed in this Feature Article.

scholarly journals An Uncertainty Estimation Framework for Probabilistic Object Detection

2021 ◽  
Author(s):  
Zongyao Lyu ◽  
Nolan B. Gutierrez ◽  
William J. Beksi
Keyword(s):  
Object Detection ◽  
Uncertainty Estimation

scholarly journals Deep Bayesian Gaussian processes for uncertainty estimation in electronic health records

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Yikuan Li ◽  
Shishir Rao ◽  
Abdelaali Hassaine ◽  
Rema Ramakrishnan ◽  
Dexter Canoy ◽  
...  
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Gaussian Processes ◽  
Large Scale ◽  
Clinical Decision Making ◽  
Uncertainty Estimation ◽  
Kernel Learning ◽  
Estimation Methods ◽  
Minority Class ◽  
Bayesian Neural Networks

AbstractOne major impediment to the wider use of deep learning for clinical decision making is the difficulty of assigning a level of confidence to model predictions. Currently, deep Bayesian neural networks and sparse Gaussian processes are the main two scalable uncertainty estimation methods. However, deep Bayesian neural networks suffer from lack of expressiveness, and more expressive models such as deep kernel learning, which is an extension of sparse Gaussian process, captures only the uncertainty from the higher-level latent space. Therefore, the deep learning model under it lacks interpretability and ignores uncertainty from the raw data. In this paper, we merge features of the deep Bayesian learning framework with deep kernel learning to leverage the strengths of both methods for a more comprehensive uncertainty estimation. Through a series of experiments on predicting the first incidence of heart failure, diabetes and depression applied to large-scale electronic medical records, we demonstrate that our method is better at capturing uncertainty than both Gaussian processes and deep Bayesian neural networks in terms of indicating data insufficiency and identifying misclassifications, with a comparable generalization performance. Furthermore, by assessing the accuracy and area under the receiver operating characteristic curve over the predictive probability, we show that our method is less susceptible to making overconfident predictions, especially for the minority class in imbalanced datasets. Finally, we demonstrate how uncertainty information derived by the model can inform risk factor analysis towards model interpretability.

scholarly journals Methane Emission Estimation of Oil and Gas Sector: A Review of Measurement Technologies, Data Analysis Methods and Uncertainty Estimation

2021 ◽  
Vol 13 (24) ◽  
pp. 13895
Author(s):  
Shuo Sun ◽  
Linwei Ma ◽  
Zheng Li
Keyword(s):  
Data Analysis ◽  
Oil And Gas ◽  
Uncertainty Estimation ◽  
Technology Selection ◽  
Estimation Methods ◽  
Emission Mitigation ◽  
Analysis Methods ◽  
Emission Estimation ◽  
Oil And Gas Sector ◽  
Data Analysis Methods

The emission estimation of the oil and gas sector, which involves field test measurements, data analysis, and uncertainty estimation, precedes effective emission mitigation actions. A systematic comparison and summary of these technologies and methods are necessary to instruct the technology selection and for uncertainty improvement, which is not found in existing literature. In this paper, we present a review of existing measuring technologies, matching data analysis methods, and newly developed probabilistic tools for uncertainty estimation and try to depict the process for emission estimation. Through a review, we find that objectives have a determinative effect on the selection of measurement technologies, matching data analysis methods, and uncertainty estimation methods. And from a systematic perspective, optical instruments may have greatly improved measurement accuracy and range, yet data analysis methods might be the main contributor of estimation uncertainty. We suggest that future studies on oil and gas methane emissions should focus on the analysis methods to narrow the uncertainty bond, and more research on uncertainty generation might also be required.

Sign in / Sign up

Export Citation Format

Share Document