scholarly journals Deep ConvNet: Non-Random Weight Initialization for Repeatable Determinism, examined with FSGM

2021 ◽  
Author(s):  
Richard Niall Mark Rudd-Orthner ◽  
Lyudmila Mihaylova
Keyword(s):  
Human Performance ◽  
Random Number ◽  
Validation Dataset ◽  
Random Weight ◽  
Random Initialization ◽  
Fast Gradient ◽  
Weight Initialization ◽  
Single Epoch ◽  
Better Than ◽  
Sign Method

This paper presents a non-random weight initialization method in convolutional layers of neural networks examined with the Fast Gradient Sign Method (FSGM) attack. This paper's focus is convolutional layers, and are the layers that have been responsible for better than human performance in image categorization. The proposed method induces earlier learning through the use of striped forms, and as such has less unlearning of the existing random number speckled methods, consistent with the intuitions of Hubel and Wiesel. The proposed method provides a higher performing accuracy in a single epoch, with improvements of between 3-5% in a well known benchmark model, of which the first epoch is the most relevant as it is the epoch after initialization. The proposed method is also repeatable and deterministic, as a desirable quality for safety critical applications in image classification within sensors. That method is robust to Glorot/Xavier and He initialization limits as well. The proposed non-random initialization was examined under adversarial perturbation attack through the FGSM approach with transferred learning, as a technique to measure the affect in transferred learning with controlled distortions, and finds that the proposed method is less compromised to the original validation dataset, with higher distorted datasets.

scholarly journals Deep ConvNet: Non-Random Weight Initialization for Repeatable Determinism, Examined with FSGM

Sensors ◽  
2021 ◽  
Vol 21 (14) ◽  
pp. 4772
Author(s):  
Richard N. M. Rudd-Orthner ◽  
Lyudmila Mihaylova
Keyword(s):  
Cross Validation ◽  
Color Image ◽  
Validation Dataset ◽  
Number Sequence ◽  
Random Weight ◽  
Fast Gradient ◽  
Weight Initialization ◽  
Fitting In ◽  
Sequence Substitution ◽  
Sign Method

A repeatable and deterministic non-random weight initialization method in convolutional layers of neural networks examined with the Fast Gradient Sign Method (FSGM). Using the FSGM approach as a technique to measure the initialization effect with controlled distortions in transferred learning, varying the dataset numerical similarity. The focus is on convolutional layers with induced earlier learning through the use of striped forms for image classification. Which provided a higher performing accuracy in the first epoch, with improvements of between 3–5% in a well known benchmark model, and also ~10% in a color image dataset (MTARSI2), using a dissimilar model architecture. The proposed method is robust to limit optimization approaches like Glorot/Xavier and He initialization. Arguably the approach is within a new category of weight initialization methods, as a number sequence substitution of random numbers, without a tether to the dataset. When examined under the FGSM approach with transferred learning, the proposed method when used with higher distortions (numerically dissimilar datasets), is less compromised against the original cross-validation dataset, at ~31% accuracy instead of ~9%. This is an indication of higher retention of the original fitting in transferred learning.

scholarly journals Implementation and Comparison of Two Pharmacometric Tools for Model-Based Therapeutic Drug Monitoring and Precision Dosing of Daptomycin

Pharmaceutics ◽  
2022 ◽  
Vol 14 (1) ◽  
pp. 114
Author(s):  
Justine Heitzmann ◽  
Yann Thoma ◽  
Romain Bricca ◽  
Marie-Claude Gagnieu ◽  
Vincent Leclerc ◽  
...  
Keyword(s):  
Therapeutic Drug Monitoring ◽  
Drug Monitoring ◽  
Joint Infection ◽  
Significant Proportion ◽  
Initial Dosage ◽  
Validation Dataset ◽  
Therapeutic Drug ◽  
Efficacy And Safety ◽  
Absolute Percent Error ◽  
Better Than

Daptomycin is a candidate for therapeutic drug monitoring (TDM). The objectives of this work were to implement and compare two pharmacometric tools for daptomycin TDM and precision dosing. A nonparametric population PK model developed from patients with bone and joint infection was implemented into the BestDose software. A published parametric model was imported into Tucuxi. We compared the performance of the two models in a validation dataset based on mean error (ME) and mean absolute percent error (MAPE) of individual predictions, estimated exposure and predicted doses necessary to achieve daptomycin efficacy and safety PK/PD targets. The BestDose model described the data very well in the learning dataset. In the validation dataset (94 patients, 264 concentrations), 21.3% of patients were underexposed (AUC24h < 666 mg.h/L) and 31.9% of patients were overexposed (Cmin > 24.3 mg/L) on the first TDM occasion. The BestDose model performed slightly better than the model in Tucuxi (ME = −0.13 ± 5.16 vs. −1.90 ± 6.99 mg/L, p < 0.001), but overall results were in agreement between the two models. A significant proportion of patients exhibited underexposure or overexposure to daptomycin after the initial dosage, which supports TDM. The two models may be useful for model-informed precision dosing.

Enhancing adversarial attack transferability with multi-scale feature attack

2020 ◽  
pp. 2050076
Author(s):  
Caixia Sun ◽  
Lian Zou ◽  
Cien Fan ◽  
Yu Shi ◽  
Yifeng Liu
Keyword(s):  
Internal Representation ◽  
Feature Space ◽  
Source Model ◽  
Black Box ◽  
Space Representation ◽  
Scale Feature ◽  
Multi Scale ◽  
Adversarial Examples ◽  
Fast Gradient ◽  
Sign Method

Deep neural networks are vulnerable to adversarial examples, which can fool models by adding carefully designed perturbations. An intriguing phenomenon is that adversarial examples often exhibit transferability, thus making black-box attacks effective in real-world applications. However, the adversarial examples generated by existing methods typically overfit the structure and feature representation of the source model, resulting in a low success rate in a black-box manner. To address this issue, we propose the multi-scale feature attack to boost attack transferability, which adjusts the internal feature space representation of the adversarial image to get far to the internal representation of the original image. We show that we can select a low-level layer and a high-level layer of the source model to conduct the perturbations, and the crafted adversarial examples are confused with original images, not just in the class but also in the feature space representations. To further improve the transferability of adversarial examples, we apply reverse cross-entropy loss to reduce the overfitting further and show that it is effective for attacking adversarially trained models with strong defensive ability. Extensive experiments show that the proposed methods consistently outperform the iterative fast gradient sign method (IFGSM) and momentum iterative fast gradient sign method (MIFGSM) under the challenging black-box setting.

scholarly journals The bounded rationality of probability distortion

2020 ◽  
Vol 117 (36) ◽  
pp. 22024-22034 ◽  
Author(s):  
Hang Zhang ◽  
Xiangjuan Ren ◽  
Laurence T. Maloney
Keyword(s):  
Bounded Rationality ◽  
Relative Frequency ◽  
Human Performance ◽  
Model Comparison ◽  
Separate Analysis ◽  
Decision Making Under Risk ◽  
Factorial Model ◽  
Frequency Judgments ◽  
Log Odds ◽  
Better Than

In decision making under risk (DMR) participants’ choices are based on probability values systematically different from those that are objectively correct. Similar systematic distortions are found in tasks involving relative frequency judgments (JRF). These distortions limit performance in a wide variety of tasks and an evident question is, Why do we systematically fail in our use of probability and relative frequency information? We propose a bounded log-odds model (BLO) of probability and relative frequency distortion based on three assumptions: 1) log-odds: probability and relative frequency are mapped to an internal log-odds scale, 2) boundedness: the range of representations of probability and relative frequency are bounded and the bounds change dynamically with task, and 3) variance compensation: the mapping compensates in part for uncertainty in probability and relative frequency values. We compared human performance in both DMR and JRF tasks to the predictions of the BLO model as well as 11 alternative models, each missing one or more of the underlying BLO assumptions (factorial model comparison). The BLO model and its assumptions proved to be superior to any of the alternatives. In a separate analysis, we found that BLO accounts for individual participants’ data better than any previous model in the DMR literature. We also found that, subject to the boundedness limitation, participants’ choice of distortion approximately maximized the mutual information between objective task-relevant values and internal values, a form of bounded rationality.

BIOLOGICALLY PLAUSIBLE BSDT RECOGNITION OF COMPLEX IMAGES: THE CASE OF HUMAN FACES

2008 ◽  
Vol 18 (06) ◽  
pp. 527-545 ◽  
Author(s):  
PETRO GOPYCH
Keyword(s):  
Signal Detection Theory ◽  
Human Performance ◽  
Recognition Performance ◽  
Behavioral Effects ◽  
Recognition Experiment ◽  
Binary Signal ◽  
Computer Codes ◽  
Human Faces ◽  
Better Than ◽  
Complex Images

On the basis of recent binary signal detection theory (BSDT), optimal recognition algorithms for complex images are constructed and their optimal performance are calculated. A methodology for comparing BSDT predictions and measured human performance is developed and applied to explaining particular face recognition experiment. The BSDT makes possible computer codes with recognition performance better than that in humans, its fundamental discreteness is consistent with the experiment. Related neurobiological and behavioral effects are briefly discussed.

SIMULATION PROCEDURES IN CONDENSED MATTER PHYSICS: AN ALGORITHM ALTERNATIVE TO THE STANDARD RANDOM NUMBER GENERATORS

1994 ◽  
Vol 08 (19) ◽  
pp. 1163-1173
Author(s):  
ALESSANDRO CORDELLI
Keyword(s):  
Random Number ◽  
Condensed Matter ◽  
Vibrational Structure ◽  
Periodic Systems ◽  
Solid State Physics ◽  
Random Number Generators ◽  
Simulation Techniques ◽  
Random Hamiltonian ◽  
Better Than

Numerical simulation techniques play a very important role in solid state physics, in particular, as far as the determination of electronic and vibrational structure of non-periodic systems is concerned. The basis of these techniques is the construction of a random Hamiltonian and a random state of interest; to do that, standard congruential algorithms for the generation of pseudorandom numbers are commonly used. The aim of this paper is to propose a novel, alternative way for the generation of random operators and vectors. This technique, based on the concept of minimum correlation sequence, gives results equivalent to or better than the standard approach.

Gender Differences in Navigation and Wayfinding using Mobile Augmented Reality

2005 ◽  
Vol 49 (21) ◽  
pp. 1868-1872 ◽  
Author(s):  
Ali M. Ahmad ◽  
Brian F. Goldiez ◽  
P. A. Hancock
Keyword(s):  
Augmented Reality ◽  
Human Performance ◽  
Search And Rescue ◽  
Critical Aspect ◽  
Experimental Conditions ◽  
On Demand ◽  
Post Test ◽  
Z Scores ◽  
Males And Females ◽  
Better Than

Augmented Reality (AR) technology is sufficiently mature, where it is possible to evaluate improvement in human performance. A critical aspect of human performance is individual differences in AR. In the present study, the effect of gender on human performance in a “search and rescue” navigation task is assessed. Six conditions were investigated in the study: Two control conditions (paper map or compass prior to entering the maze), and four experimental conditions (combinations of egocentric and exocentric maps, and a continuously-on or on-demand map display). 120 subjects equally divided between males and females were tested. Pre and post test questionnaires were administered. Guilford-Zimmerman (G-Z) scores indicate that males perform better than females in spatial visualization and orientation tasks. The time for maze traversal for females exceeded that of males by 127 seconds on average for the no map condition. Also, males had better performance in covering the maze.

Sign in / Sign up

Export Citation Format

Share Document