On the end-performance metric estimator selection

Automatica ◽  
2015 ◽  
Vol 58 ◽  
pp. 22-27
Author(s):  
Dimitrios Katselis ◽  
Cristian R. Rojas ◽  
Boris I. Godoy ◽  
Juan C. Agüero ◽  
Carolyn L. Beck
Keyword(s):  
Performance Metric ◽  
Estimator Selection

scholarly journals Estimator selection: End-performance metric aspects

2015 ◽  
Author(s):  
Dimitrios Katselis ◽  
Cristian R. Rojas ◽  
Carolyn L. Beck
Keyword(s):  
Performance Metric ◽  
Estimator Selection

scholarly journals Stress testing reveals gaps in clinic readiness of image-based diagnostic artificial intelligence models

2021 ◽  
Vol 4 (1) ◽  
Author(s):  
Albert T. Young ◽  
Kristen Fernandez ◽  
Jacob Pfau ◽  
Rasika Reddy ◽  
Nhat Anh Cao ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Real World ◽  
Stress Testing ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Skin Lesions ◽  
Stress Tests ◽  
Performance Metric ◽  
Intelligence Models ◽  
Artificial Intelligence Models

AbstractArtificial intelligence models match or exceed dermatologists in melanoma image classification. Less is known about their robustness against real-world variations, and clinicians may incorrectly assume that a model with an acceptable area under the receiver operating characteristic curve or related performance metric is ready for clinical use. Here, we systematically assessed the performance of dermatologist-level convolutional neural networks (CNNs) on real-world non-curated images by applying computational “stress tests”. Our goal was to create a proxy environment in which to comprehensively test the generalizability of off-the-shelf CNNs developed without training or evaluation protocols specific to individual clinics. We found inconsistent predictions on images captured repeatedly in the same setting or subjected to simple transformations (e.g., rotation). Such transformations resulted in false positive or negative predictions for 6.5–22% of skin lesions across test datasets. Our findings indicate that models meeting conventionally reported metrics need further validation with computational stress tests to assess clinic readiness.

Characteristics of Long-run Return and Risk: A Unified Performance Metric

2017 ◽  
Vol 39 (2) ◽  
pp. 165-188
Author(s):  
Ping Cheng ◽  
Zhenguo Lin ◽  
Yingchun Liu
Keyword(s):  
Performance Metric ◽  
Long Run

Development of Team Coordination and Performance Measures in a Trauma Setting

2007 ◽  
Vol 51 (11) ◽  
pp. 717-721 ◽  
Author(s):  
Stuart Marshall ◽  
Anne Miller ◽  
Yan Xiao
Keyword(s):  
Team Performance ◽  
Decision Support Tool ◽  
Stage Model ◽  
Trauma Resuscitation ◽  
Team Coordination ◽  
Support Tool ◽  
University Of Maryland ◽  
Performance Metric ◽  
And Performance ◽  
The University

The paucity of reliable measures of team coordination and performance significantly obstructs the assessment of the effects of any technology on teams to improve decision making in health care. A pilot study was conducted to determine if measures of coordination and performance could be developed for teams involved in trauma resuscitation. A video assisted review of cases enabled evaluation of the use of the tools. Descriptors of coordination were derived from Klein's five-stage model of team coordination. A scoring system of team performance was developed from the University of Maryland Team Observable Performance Metric (UMTOP). After some modification both coordination and performance could be described. However, four defined stages of resuscitation were observed which greatly improved coding. More rigorous assessments of these tools will be required before firm conclusions can be drawn about the effects of a decision support tool recently introduced into the environment.

Parameter Characterization for Elastic Energy Absorption of an Embedded Corrugated Wave Padding Concept

2016 ◽  
Author(s):  
Sean S. Tolman ◽  
Amanda Beatty ◽  
Anton E. Bowden ◽  
Larry L. Howell
Keyword(s):  
Energy Absorption ◽  
Cross Section ◽  
Peak Pressure ◽  
Force Distribution ◽  
Finite Element Analyses ◽  
Performance Metric ◽  
Physical Testing ◽  
Physical Tests ◽  
Factorial Design Of Experiments ◽  
Full Factorial

The parameters of an innovative padding concept were investigated using Finite Element Analyses (FEA) and physical testing. The concept relies on a compliant corrugation embedded in an elastic foam to provide stiffness for force distribution and elastic deformation for energy absorption. The shape of the corrugation cross section was explored as well as the wavelength and amplitude by employing a full factorial design of experiments. FEA results were used to choose designs for prototyping and physical testing. The results of the physical tests were consistent with the FEA predictions although the FEA tended to underestimate the peak pressure compared to the physical tests. A performance metric is proposed to compare different padding configurations. The concept shows promise for sports padding applications. It may allow for designs which are smaller, more lightweight, and move better with an athlete than current technologies yet still provide the necessary protective functions.

Application-Oriented Estimator Selection

2015 ◽  
Vol 22 (4) ◽  
pp. 489-493
Author(s):  
Dimitrios Katselis ◽  
Cristian R. Rojas
Keyword(s):  
Estimator Selection

US2RO: Union of Superpoints to Recognize Objects

2021 ◽  
Vol 15 (04) ◽  
pp. 513-537
Author(s):  
Marcel Tiator ◽  
Anna Maria Kerkmann ◽  
Christian Geiger ◽  
Paul Grimm
Keyword(s):  
Imitation Learning ◽  
Network Architectures ◽  
Application Development ◽  
Repetitive Work ◽  
Data Set ◽  
Partition System ◽  
Performance Metric ◽  
Growing Environment ◽  
The Creation ◽  
Near Future

The creation of interactive virtual reality (VR) applications from 3D scanned content usually includes a lot of manual and repetitive work. Our research aim is to develop agents that recognize objects to enhance the creation of interactive VR applications. We trained partition agents in our superpoint growing environment that we extended with an expert function. This expert function solves the sparse reward signal problem of the previous approaches and enables to use a variant of imitation learning and deep reinforcement learning with dense feedback. Additionally, the function allows to calculate a performance metric for the degree of imitation for different partitions. Furthermore, we introduce an environment to optimize the superpoint generation. We trained our agents with 1182 scenes of the ScanNet data set. More specifically, we trained different neural network architectures with 1170 scenes and tested their performance with 12 scenes. Our intermediate results are promising such that our partition system might be able to assist the VR application development from 3D scanned content in near future.

Hybrid Beamforming for DFRC System Based on SINR Performance Metric

2021 ◽  
Author(s):  
Yuanzhe Dai ◽  
Kaifeng Han ◽  
Guo Wei ◽  
Ying Du ◽  
Zhiqin Wang
Keyword(s):  
Performance Metric ◽  
Hybrid Beamforming
Sign in / Sign up

Export Citation Format

Share Document