Search

Deep neural networks have exhibited remarkable performance in various domains. However, the reliance of these models on spurious features has raised concerns about their reliability. A promising solution to this problem is last-layer retraining, which involves retraining the linear classifier head on a small subset of data without spurious cues. Nevertheless, selecting this subset requires human supervision, which reduces its scalability. Moreover, spurious cues may still exist in the selected subset. As a solution to this problem, we propose a novel ranking framework that leverages an open vocabulary object detection technique to identify images without spurious cues. More specifically, we use the object detector as a measure to score the presence of the target object in the images. Next, the images are sorted based on this score, and the last-layer of the model is retrained on a subset of the data with the highest scores. Our experiments on the ImageNet-1k dataset demonstrate the effectiveness of this ranking framework in sorting images based on spuriousness and using them for last-layer retraining., Comment: Accepted at ICML 2023 Workshop on Spurious Correlations, Invariance, and Stability (SCIS)

Despite the remarkable success achieved by deep learning algorithms in various domains, such as computer vision, they remain vulnerable to adversarial perturbations. Adversarial Training (AT) stands out as one of the most effective solutions to address this issue; however, single-step AT can lead to Catastrophic Overfitting (CO). This scenario occurs when the adversarially trained network suddenly loses robustness against multi-step attacks like Projected Gradient Descent (PGD). Although several approaches have been proposed to address this problem in Convolutional Neural Networks (CNNs), we found out that they do not perform well when applied to Vision Transformers (ViTs). In this paper, we propose Blacksmith, a novel training strategy to overcome the CO problem, specifically in ViTs. Our approach utilizes either of PGD-2 or Fast Gradient Sign Method (FGSM) randomly in a mini-batch during the adversarial training of the neural network. This will increase the diversity of our training attacks, which could potentially mitigate the CO issue. To manage the increased training time resulting from this combination, we craft the PGD-2 attack based on only the first half of the layers, while FGSM is applied end-to-end. Through our experiments, we demonstrate that our novel method effectively prevents CO, achieves PGD-2 level performance, and outperforms other existing techniques including N-FGSM, which is the state-of-the-art method in fast training for CNNs.

The purpose of this study was to examine the effect of job stress on the commitment and turnover of the employees of Saderat Bank in Zabol by considering the mediating role of job burnout. The statistical population of this study consists of all employees and managers of branches of Saderat Bank in Zabol, who are currently working. According to the prosecution, the size of the population is estimated 1248 people. For sampling, the Cochran sample size formula with limited population was used. Finally, 295 statistical units were analyzed. The data collection instrument used in this study was a series of standard questionnaires. Additionally, in order to describe the data analysis and to test the hypothesis of the research, inferential statistics and structural equation modeling method have been usedand analyzed via LISREL. After analyzing the data, the results of research hypothesis testing indicate that job stress has a positive effect on job burnout and job stress has positive effect on turnover. In addition, the negative effect of job burnout on organizational commitment was accepted.Keywords: Job Stress, Job Burnout, Organizational Commitment, Branches of Saderat Bank in Zabol.

 Fuzzy logic is a tool to use human expertise. The simplicity of fuzzy-rule based systems and its power to perform various tasks without accurate measurement and computation makes it very popular between sciences. One of these applications is using fuzzy logic in designing the controller for navigation of autonomous robots to move in various environments. This paper proposes a new method of robot navigation based on fuzzy logic. This method can be drawn upon to design robots which can find and catch different kind of animals, especially endangered species. It works based on a hierarchical set of behavior each of which acts by using a set of fuzzy rules. The proposed method is simulated and tested by MATLAB software.

It is still not entirely clear how α-galactosidase A (GAA) deficiency translates into clinical symptoms of Fabry disease (FD). The present communication investigates the effects of the mutation N215S in FD on the trafficking and processing of lysosomal GAA and their potential association with alterations in the membrane lipid composition. Abnormalities in lipid rafts (LRs) were observed in fibroblasts isolated from a male patient with FD bearing the mutation N215S. Interestingly, LR analysis revealed that the distribution of cholesterol and flotillin-2 are distinctly altered in the Fabry fibroblasts when compared with that of the wild-type cells. Furthermore, increased levels of glycolipid globotriaosylceramide 3 (Gb3) and sphingomyelin (SM) were observed in non-raft membrane fractions of Fabry cells. Substrate reduction with N-butyldeoxynojirimycin (NB-DNJ) in vitro was capable of reversing these abnormalities in this patient. These data led to the hypothesis that alterations of LRs may contribute to the pathophysiology of Morbus Fabry. Furthermore, it may be suggested that substrate reduction therapy with NB-DNJ might be a promising approach for the treatment of GAA deficiency at least for the selected patients.