Image Complexity Guided Network Compression for Biomedical Image Segmentation

Compression is a standard procedure for making convolutional neural networks (CNNs) adhere to some specific computing resource constraints. However, searching for a compressed architecture typically involves a series of time-consuming training/validation experiments to determine a good compromise between network size and performance accuracy. To address this, we propose an image complexity-guided network compression technique for biomedical image segmentation. Given any resource constraints, our framework utilizes data complexity and network architecture to quickly estimate a compressed model which does not require network training. Specifically, we map the dataset complexity to the target network accuracy degradation caused by compression. Such mapping enables us to predict the final accuracy for different network sizes, based on the computed dataset complexity. Thus, one may choose a solution that meets both the network size and segmentation accuracy requirements. Finally, the mapping is used to determine the convolutional layer-wise multiplicative factor for generating a compressed network. We conduct experiments using 5 datasets, employing 3 commonly-used CNN architectures for biomedical image segmentation as representative networks. Our proposed framework is shown to be effective for generating compressed segmentation networks, retaining up to ≈95% of the full-sized network segmentation accuracy, and at the same time, utilizing ≈32x fewer network trainable weights (average reduction) of the full-sized networks.

A hierarchical RCNN for vehicle and vehicle license plate detection and recognition

<span>Vehicle and vehicle license detection obtained incredible achievements during recent years that are also popularly used in real traffic scenarios, such as intelligent traffic monitoring systems, auto parking systems, and vehicle services. Computer vision attracted much attention in vehicle and vehicle license detection, benefit from image processing and machine learning technologies. However, the existing methods still have some issues with vehicle and vehicle license plate recognition, especially in a complex environment. In this paper, we propose a multivehicle detection and license plate recognition system based on a hierarchical region convolutional neural network (RCNN). Firstly, a higher level of RCNN is employed to extract vehicles from the original images or video frames. Secondly, the regions of the detected vehicles are input to a lower level (smaller) RCNN to detect the license plate. Thirdly, the detected license plate is split into single numbers. Finally, the individual numbers are recognized by an even smaller RCNN. The experiments on the real traffic database validated the proposed method. Compared with the commonly used all-in-one deep learning structure, the proposed hierarchical method deals with the license plate recognition task in multiple levels for sub-tasks, which enables the modification of network size and structure according to the complexity of sub-tasks. Therefore, the computation load is reduced.</span>

The link between HRD professionals’ social capital and career adaptability: a moderated mediation analysis of social network

Purpose From a social capital perspective, this study aims to shed light on the link between social capital and career adaptability by focusing on how social connections and interactions shape and nurture career adaptability. Drawing on socioemotional selectivity theory, the authors further examined the critical moderating role of age on the above relationship. Design/methodology/approach Survey responses from 208 HRD professionals were analyzed via a moderated mediation analysis. Findings The results showed that there is a positive relationship between social capital (network size and intimate network) and career adaptability; frequent interaction increases intimacy, in turn enhancing career adaptability; and the indirect effect of social capital on career adaptability (via intimate network) is stronger when the employee is younger. Originality/value The most novel theoretical contribution of this study is that the authors lend empirical support to the connection between social capital and career adaptability moderated by age. The study also contributes to understanding how core aspects of social capital are inter-related each other and have directional relationships.

The Social Networks and Social Support of Siblings of Children with Cancer

Siblings of children with cancer need support to ameliorate the challenges they encounter; however, little is known about what types and sources of support exist for siblings. This study addresses this gap in our understanding of the social networks and sources of support for adolescents with a brother or sister who has cancer. Additionally, we describe how the support siblings receive addresses what they feel are the hardest aspects of being a sibling of a child with cancer. During semi-structured interviews, siblings (ages 12–17) constructed ecomaps describing their support networks. Data were coded for support type (emotional, instrumental, informational, validation, companionship) and support provider (e.g., mother, teacher, friend). Network characteristics and patterns of support were explored. Support network size ranged from 3 to 10 individuals (M = 6 ± 1.9); siblings most frequently reported mothers as sources of support (n = 22, 91.7%), followed by fathers (n = 19, 79.2%), close friends (n = 19, 79.2%) and siblings (with or without cancer) (n = 17, 70.8%). Friends and brothers or sisters most often provided validation and companionship while instrumental and informational supports came from parents. This study provides foundational knowledge about siblings’ support networks, which can be utilized to design interventions that improve support for siblings of children with cancer.

Dispositional optimism and depression risk in older women in the Nurses´ Health Study: a prospective cohort study

Dispositional optimism is a potentially modifiable factor and has been associated with multiple physical health outcomes, but its relationship with depression, especially later in life, remains unclear. In the Nurses´ Health Study (n = 33,483), we examined associations between dispositional optimism and depression risk in women aged 57–85 (mean = 69.9, SD = 6.8), with 4,051 cases of incident depression and 10 years of follow-up (2004–2014). We defined depression as either having a physician/clinician-diagnosed depression, or regularly using antidepressants, or the presence of severe depressive symptoms using validated self-reported scales. Age- and multivariable-adjusted Cox proportional hazards models were used to estimate hazard ratios (HRs) with 95% confidence intervals (95% CIs) across optimism quartiles and for a 1-standard deviation (SD) increment of the optimism score. In sensitivity analyses we explored more restrictive definitions of depression, potential mediators, and moderators. In multivariable-adjusted models, women with greater optimism (top vs. bottom quartile) had a 27% (95%CI = 19–34%) lower risk of depression. Every 1-SD increase in the optimism score was associated with a 15% (95%CI = 12–18%) lower depression risk. When applying a more restrictive definition for clinical depression, the association was considerably attenuated (every 1-SD increase in the optimism score was associated with a 6% (95%CI = 2–10%-) lower depression risk. Stratified analyses by baseline depressive symptoms, age, race, and birth region revealed comparable estimates, while mediators (emotional support, social network size, healthy lifestyle), when combined, explained approximately 10% of the optimism-depression association. As social and behavioral factors only explained a small proportion of the association, future research should investigate other potential pathways, such as coping strategies, that may relate optimism to depression risk.

Revisiting Extraversion and Leadership Emergence: A Social Network Churn Perspective

One of the classic relationships in personality psychology is that extraversion is associated with emerging as an informal leader. However, recent findings raise questions about the longevity of extraverted individuals as emergent leaders. Here, we adopt a social network churn perspective to study the number of people entering, remaining in, and leaving the leadership networks of individuals over time. We propose that extraverted individuals endure as emergent leaders in networks over time, but experience significant changes in the people being led, including the loss of people who once considered them a leader but now no longer do. In Study 1 (N = 545), extraverted individuals had a larger number of new and remaining people in their leadership networks, but also lost more people, above and beyond differences in initial leadership network size. In Study 2 (N = 764), we replicated and extended these results in an organizational sample while controlling for alternative explanations such as formal rank, network size, self-monitoring, and narcissism. Extraversion predicted the number of people entering, remaining in, and leaving leadership networks over time. Our findings suggest that while extraverted individuals tend to emerge as leaders, they are also more likely to experience greater network churn—they tend to lead different people over time and leave people in their wake who once perceived them a leader but now no longer do. We discuss the challenges posed by this network churn perspective for extraverted emergent leaders and highlight its importance for our understanding of extraversion and emergent leadership.

A Novel High-Capacity Information Hiding Scheme Based on Improved U-Net

With the gradual introduction of deep learning into the field of information hiding, the capacity of information hiding has been greatly improved. Therefore, a solution with a higher capacity and a good visual effect had become the current research goal. A novel high-capacity information hiding scheme based on improved U-Net was proposed in this paper, which combined improved U-Net network and multiscale image analysis to carry out high-capacity information hiding. The proposed improved U-Net structure had a smaller network scale and could be used in both information hiding and information extraction. In the information hiding network, the secret image was decomposed into wavelet components through wavelet transform, and the wavelet components were hidden into image. In the extraction network, the features of the hidden image were extracted into four components, and the extracted secret image was obtained. Both the hiding network and the extraction network of this scheme used the improved U-Net structure, which preserved the details of the carrier image and the secret image to the greatest extent. The simulation experiment had shown that the capacity of this scheme was greatly improved than that of the traditional scheme, and the visual effect was good. And compared with the existing similar solution, the network size has been reduced by nearly 60%, and the processing speed has been increased by 20%. The image effect after hiding the information was improved, and the PSNR between the secret image and the extracted image was improved by 6.3 dB.

Per-link Parallel and Distributed Hybrid Beamforming for Multi-Cell Massive MIMO Millimeter Wave Full Duplex

This paper presents two novel hybrid beamforming (HYBF) designs for a multi-cell massive multiple-input-multiple-output (mMIMO) millimeter wave (mmWave) full duplex (FD) system under limited dynamic range (LDR). Firstly, we present a novel centralized HYBF (C-HYBF) scheme based on alternating optimization. In general, the complexity of C-HYBF schemes scales quadratically as a function of the number of users and cells, which may limit their scalability. Moreover, they require significant communication overhead to transfer complete channel state information (CSI) to the central node every channel coherence time for optimization. The central node also requires very high computational power to jointly optimize many variables for the uplink (UL) and downlink (DL) users in FD systems. To overcome these drawbacks, we propose a very low-complexity and scalable cooperative per-link parallel and distributed (P$\&$D)-HYBF scheme. It allows each mmWave FD base station (BS) to update the beamformers for its users in a distributed fashion and independently in parallel on different computational processors. The complexity of P$\&$D-HYBF scales only linearly as the network size grows, making it desirable for the next generation of large and dense mmWave FD networks. Simulation results show that both designs significantly outperform the fully digital half duplex (HD) system with only a few radio-frequency (RF) chains and achieve similar performance. <br>

Hardware Architecture for Asynchronous Cellular Self-Organizing Maps

Nowadays, one of the main challenges in computer architectures is scalability; indeed, novel processor architectures can include thousands of processing elements on a single chip and using them efficiently remains a big issue. An interesting source of inspiration for handling scalability is the mammalian brain and different works on neuromorphic computation have attempted to address this question. The Self-configurable 3D Cellular Adaptive Platform (SCALP) has been designed with the goal of prototyping such types of systems and has led to the proposal of the Cellular Self-Organizing Maps (CSOM) algorithm. In this paper, we present a hardware architecture for CSOM in the form of interconnected neural units with the specific property of supporting an asynchronous deployment on a multi-FPGA 3D array. The Asynchronous CSOM (ACSOM) algorithm exploits the underlying Network-on-Chip structure to be provided by SCALP in order to overcome the multi-path propagation issue presented by a straightforward CSOM implementation. We explore its behaviour under different map topologies and scalar representations. The results suggest that a larger network size with low precision coding obtains an optimal ratio between algorithm accuracy and FPGA resources.

Initial social network analysis of producers working towards sustainability suggests weak ties and potential fragmentation

For farmers to adopt and maintain sustainable farming practices, they must have the resources and network to succeed with this work and must realize a positive impact on their business model. As a food system is ultimately made up of the people, organizations, and institutions that grow, move, buy and sell food, we must understand who is at the center of this network, who is well-connected, and who is peripheral. Within a particular regional food system in a highly productive southeastern U.S. state, the network of local producers interested in sustainable production, including environmental and economic components, seems to be growing. However, it is unclear who benefits from this system and whether this system is growing in a way that encourages and enhances the benefits for sustainable agriculture. Existing evidence for the network size and its vulnerabilities has been anecdotal, from Extension agents and their contacts with individual producers, rather than based on systematic research. We used social network analysis to understand the status of the system and its constituents. Connections between producers appear to be weak overall with potential fragmentation, suggesting a fragility that could easily derail efforts to increase sustainable production in the region.