投桃报李:临时员工助人行为对正式员工隐性知识分享的影响机制

多元雇佣触发的员工内部关系矛盾问题日趋复杂,如何促进多元雇佣聚力发展,已成为零工经济情境下理论和实践领域着力探讨的核心议题。文章基于社会交换理论,引入关系认同和感知义务作为中介变量,将任务相依性作为边界条件,构建临时员工助人行为对正式员工隐性知识分享的双路径机理模型,旨在分析临时员工助人行为对正式员工隐性知识分享的作用效果及潜在机制。通过对三阶段收集的237份纵向数据分析表明:关系认同和感知义务在临时员工助人行为与正式员工隐性知识分享之间起中介作用;任务相依性会强化临时员工助人行为对关系认同和感知义务的影响效应,并且任务相依性能够调节关系认同和感知义务的中介路径,即任务相依性程度越强,临时员工通过关系认同和感知义务对正式员工隐性知识分享的正向影响越强。

混合现实中基于GPU虚拟化的AI计算优化

研究探讨混合现实(MR)应用中,通过GPU虚拟化优化AI计算,聚焦于多任务调度与资源共享。研究提出了一个模型,其包含一种根据任务优先级、资源需求和等待时间,动态为正在执行的任务分配GPU资源的机制。同时,模型采用优化的多任务调度算法,以提高调度效率。实验结果表明,尽管在单任务性能测试中模型的执行时间、GPU利用率和内存使用方面略逊于物理GPU,但在多任务并发和资源共享方面,研究提出的模型展现了显著优势。未来研究将探索设计更公平高效的资源共享策略,以及进一步优化多任务调度算法。

联合语义分割的自监督单目深度估计方法

研究深度估计和语义分割的图像之间的互利关系,提出了一种联合语义分割的自监督单目深度估计方法USegDepth.语义分割和深度估计任务通过共享编码器,实现语义引导.为了进一步提高编码器的跨多任务性能,设计了多任务特征提取模块,堆叠该模块构成共享编码器,解决有限感受野和缺乏跨通道交互导致的模型特征表示能力欠佳问题,进一步提升模型精度.同时,提出跨任务交互模块,通过双向的跨域信息交互细化特征表示,提升深度估计表现,特别是光度一致性监督有限的弱纹理区域和物体边界.通过在KITTI数据集上的训练和全面评估,实验结果显示所提的USegDepth模型方法的均方相对误差相比于SGDepth降低了0.176个百分点,在阈值为1.253的阈值精度达到了98.4%,证明了USegDepth在深度预测上具有较高的准确率.

油田海陆电子文档共享系统的设计与实现

本文提出一种在多个独立部署的FileNet系统间的文档共享功能的设计思想。基于内容引擎管理方法下结合异步定时任务通过API,在海洋石油企业内让海上作业平台以更高效更便捷的方式获取陆地系统库内的共享文档。

跨领域推荐方法研究综述

近年来,随着信息技术的迅速发展的爆发性增长,这一爆发式增长推动了跨领域推荐系统的出现和发展。跨领域推荐系统的设计和实现面临着诸多挑战,包括数据异构性、领域知识融合等问题。因此,书写跨领域推荐方法的研究变得尤为重要。这些方法旨在有效地整合来自不同领域的数据和信息,同时保持推荐系统的高效性和准确性。为实现这一目标,研究者们提出了各种跨领域推荐方法,包括基于迁移学习方法、基于多任务学习的方法等跨领域推荐方法,文章将从处理步骤及优缺点梳理各跨领域推荐系统方法。

Task sharing in psychotherapy as a viable global mental health approach in resource-poor countries and also in high-resource settings

A global disparity exists between the burden of mental ill-health and the resources available to address this issue.Common mental disorders,such as mild-to-moderate depression,anxiety disorders and post-traumatic stress disorder,are associated with a significant decline in health and functioning and contribute significantly to the overall burden of disease.In view of a shortage of mental health professionals in most low-and middle-income countries,task sharing(task shifting)can make an important contribution to improving access to mental health services.Task sharing involves non-specialists in the delivery of health care.The use of lay health workers is premised on the assumption that the shorter training required for non professionals will allow an increase in the provision of required care without the need for an increase in resources and will strengthen and scale-up sustainable health systems.The need for low-tech support in the management of conunon mental disorders is widespread,and the core resource is humans trained to identify and treat those requiring mental health care.There is a strong evidence base for the effectiveness of task sharing in psychotherapy,and talk therapies,e.g.cognitive behavioral therapy,can easily be delegated to non-specialists,including non-specialist health workers,peer helpers and lay people.A(cost-)effective model of mental health care in low-resource settings as well as in high-income countries may include the medical school-based training of some psychiatrists and clinical psychologists in the treatment of referred patients with serious mental illness.These professional mental health specialists would also be trainers and supervisors of non-specialist health workers who would identify common mental health disorders and provide psychotherapy to people with these conditions in a community-based setting.Community-based generalism including task sharing rather than centralized specialism may be capable of preventing and treating many common mental disorders with minor financial expenditure.

Task sharing: situation analysis about capacity and resources for management of diabetic retinopathy in Pakistan

AIM: To assess the capacity and resources of eye care facilities, using the WHO Health Systems Framework, to manage diabetic retinopathy(DR) through task sharing.METHODS: Using purposive sampling, four participants(administrators) from four selected hospitals in two provinces in Pakistan were recruited for this cross-sectional study. A survey, to assess the capacity and resources of the selected eye care facilities for the feasibility to adopt task sharing in management of DR to prevent vision loss, was emailed to participants who were asked to complete. Responses to open-ended questions were entered into a Microsoft Excel spread sheet and inductive approach was applied for analysis.RESULTS: All the surveyed eye care facilities offer eye care services for people with diabetes and DR. All surveyed eye care facilities have a shortage in the number of human resources across all cadres. Optometrists and mid-level eye care workers did not have active roles in DR screening and management in all four hospitals. All the hospitals surveyed did not have a computerized record management system for patients who visit ophthalmologists for eye examinations. Equipment for detection and management of DR were short in number and main users were ophthalmologists. There was no policy for population-based screening program for detection of DR in any of the surveyed hospitals.CONCLUSION: A system-based approach to manage DR is needed. The capacity of eye care facilities and the potential to improve access of people with diabetes to eyecare services can be enhanced through implementation of task sharing.

A Distributed Algorithm for Parallel Multi-task Allocation Based on Profit Sharing Learning

经由联盟形成的任务分配是在多代理人系统(妈) 的几应用程序域的基本研究挑战，例如资源分配，灾难反应管理等等。怎么以一种分布式的方式分配许多未解决的任务到一些代理人，主要处理。在这篇论文，我们在自我组织、自我学习的代理人之中建议一个分布式的平行多工分配算法。处理状况，我们在二维的房间地理上驱散代理人和任务，然后介绍为寻找它的任务由的一个单个代理人的分享学习的利润(PSL ) 不断自我学习。我们也在代理人之中为通讯和协商介绍策略分配真实工作量到每个 tasked 代理人。最后，评估建议算法的有效性，我们把它与 Shehory 和 Krau 被许多研究人员在最近的年里讨论的分布式的任务分配算法作比较。试验性的结果证明建议算法罐头快速为每项任务形成一个解决的联盟。而且，建议算法罐头明确地告诉我们每个 tasked 代理人的真实工作量，并且能因此为实际控制任务提供一本特定、重要的参考书。

基于Greenplum的铁路数据共享平台设计

实现铁路既有信息系统数据共享,加快数据集成应用开发,是我国铁路信息化建设继续深入推进的必由之路。文章针对铁路企业既有数据资产来源广、数据源类型多种多样等特点,基于分布式数据仓库Greenplum,开发了适用于集中存储和管理铁路企业多源异构数据的数据共享平台。平台提供数据目录管理、数据采集、数据共享、运行监控管理与数据安全防护等功能,实现对铁路多源异构数据的自动采集、处理及规范化集中存储。平台采用高可靠、易扩展的架构设计及可视化技术,为平台管理员提供简便易用的可视化配置与运行管理工具,为铁路企业应用开发单位提供便捷高效、稳定可靠、安全合规的数据共享服务,较好地满足铁路企业数据集成应用开发的需要,也为铁路企业大数据应用构建起坚实的基座。该平台已在中国铁路兰州局集团有限公司启动试运行,创建了企业既有信息系统数据资产目录,已采集76种业务数据,为十多个信息系统提供稳定、高效的数据共享服务。

Effectiveness of Task Sharing and Task Shifting on the Uptake of Family Planning in Kenya

Background: Contraception is an inexpensive and cost-effective intervention, but health workforce shortages and restrictive policies on the roles of mid and lower-level cadres limit access to effective contraceptive methods in many settings. Task sharing and task shifting are strategies that can be adopted to increase uptake of health services including family planning. Methods: We collected data through online survey, key informant interviews and focused grouped discussions with an intervention group and that implemented the task sharing and task shifting policy guidelines and a control group that did not implement the policy. A total of 434 questionnaires were filled by health workers’ in primary health care facilities to assess effectiveness of task sharing and task shifting on the uptake of family planning services including its strengths and challenges. The questionnaire was designed with the aim of getting data on services provided by the cadres on effectiveness (number of clients, increase in use of methods, access to services), how they perceive these tasks, the bottlenecks and facilitating factors associated with the practice of task sharing and task shifting. Results: We found out that the task sharing and task shifting can expand and increase access to services as stated by 95% of the respondents. Most community health workers provided more of the family planning services at 45% with only 5% of the services of family planning being provided by medical officers. 98% of family planning services were integrated with other services. Task shifting was beneficial to the health care providers as well as the clients and the success of task sharing and task shifting depended on training, supportive supervision and a regulated environment through policies. Conclusion: The study shows that formalized task sharing and task shifting can increase health service uptake especially when community health workers are involved to provide services in the community. This leads to increased service provision, equivalent health professional performance across cadres and patient outcomes in the provision of family planning services.

Task Co-Representation in Aging: An Event-Related Potential Study

The goal of the present study was to investigate age-related changes in attentional allocation for shared task representations during joint performance;event-related potentials were recorded while participants performed a modified visual three-stimulus oddball task, both alone and together with another participant. Younger adults and older adults (14 each) participated in the study. Participants were required to identify rare target stimuli while ignoring frequent standards, as well as rare non-targets assigned to a partner’s action (i.e., no-go stimuli for one’s own task). ERP component, nogo-P3 and P3b were measured to investigate the inhibition and the attentional allocation to the partner’s stimuli. Results showed that younger adults elicited larger frontal nogo P3 and parietal P3b for non-targets in the joint than in the individual condition. Contrary to expectation, older adults induced frontal no-go P3 in the joint condition not in the individual condition. In the sharing of the task with another, the result suggested that the efficiency of matching of incoming information with the representation of the other’s task declined with age, whereas aging did not affect the suppression of incorrect preparation of motor responses instigated by this representation.

Roles of the eye care workforce for task sharing in management of diabetic retinopathy in Cambodia

AIM： To identify the current roles of eye and health care workers in eye care delivery and investigate their potential roles in screening and detection for management of diabetic retinopathy（DR） through task sharing.METHODS： Purposive sampling of 24 participants including health administrators, members from non-government organizations and all available eye care workers in Takeo province were recruited. This cross sectional mixed method study comprised a survey and in-depth interviews. Data were collected from medical records at Caritas Takeo Eye Hospital（CTEH） and Kiri Vong District Referral Hospital Vision Centre, and a survey and interviews with participants were done to explore the potential roles for task sharing in DR management. Qualitative data were transcribed into a text program and then entered into N-Vivo（version 10） software for data management and analysis. RESULTS： From 2009 to 2012, a total of 105 178 patients were examined and 14 030 eye surgeries were performed in CTEH by three ophthalmologists supported by ophthalmic nurses in operating and eye examination for patients. Between January 2011 and September 2012, 151 patients（72 males） with retinal pathology including 125（83%） with DR visited CTEH. In addition 170 patients with diabetes were referred to CTEH for eye examinations from Mo Po Tsyo screening programs for people with diabetes. Factors favouring task sharing included high demand for eye care services and scarcity of ophthalmologists.CONCLUSION： Task sharing and team work for eye care services is functional. Participants favor the potential role of ophthalmic nurses in screening for DR through task sharing.

Task scheduling scheme by checkpoint sharing and task duplication in P2P-based desktop grids

A scheduling scheme is proposed to reduce execution time by means of both checkpoint sharing and task duplication under a peer-to-peer(P2P) architecture. In the scheme, the checkpoint executed by each peer(i.e., a resource) is used as an intermediate result and executed in other peers via its duplication and transmission. As the checkpoint is close to a final result, the reduction of execution time for each task becomes higher, leading to reducing turnaround time. To evaluate the performance of our scheduling scheme in terms of transmission cost and execution time, an analytical model with an embedded Markov chain is presented. We also conduct simulations with a failure rate of tasks and compare the performance of our scheduling scheme with that of the existing scheme based on client-server architecture. Performance results show that our scheduling scheme is superior to the existing scheme with respect to the reduction of execution time and turnaround time.

基于软参数共享的事件联合抽取方法

事件抽取是项重要的信息抽取任务,旨在抽取文本中的事件信息。目前基于多任务学习的事件联合抽取方法大多基于硬参数共享,此类方法往往会导致跷跷板现象的出现,即一项任务的性能往往通过损害另一项任务的性能来提高。为了解决这一问题,提出了一种基于软参数共享的事件联合抽取方法,该方法明确地分离了共享参数和任务特定参数,并通过双层门控网络增强模型提取和筛选语义知识的能力,使模型能同时为两个任务学习到合适的特征表示,实现了更高效的信息共享和联合表示学习。在DuEE1.0公共数据集上进行了实验,使用准确率、召回率、F_(1)值作为评价指标,并通过对比实验和消融实验验证了方法的有效性。对比基于硬参数共享的联合抽取模型事件识别任务F_(1)值提高了2.0%,论元角色分类任务F_(1)值提高了0.9%,有效地缓解了跷跷板现象的出现,验证了方法的有效性。

政府补贴和第三方回收下的产品回收合作策略

为探讨第三方回收模式下供应链成员的回收合作策略问题,构建由制造商、零售商及第三方回收商组成的供应链系统,以成本分担和任务分担两种主要的回收合作模式为基础,研究政府给予制造商补贴时,制造商和回收商的回收合作策略,对比不合作及两种合作模式下供应链成员的最优决策及利润,并结合制造商与回收商的回收成本系数分析政府补贴对供应链各成员利润的影响.研究表明:政府补贴对供应链成员利润有正向影响,无论政府补贴怎样变化,制造商与回收商合作时的产品回收率、供应链成员利润及其随补贴的增长幅度总是高于不合作时;当政府补贴、制造商回收成本系数等满足一定条件时,在成本分担模式下有更大回收率,且供应链成员能够获得更大利润.

云计算平台下政务大数据信息资源共享模型构建

政务大数据共享模型链路干扰较大,存在安全信道构建失败的情况,为此提出云计算平台下政务大数据信息资源共享模型。将待共享资源上传到云计算平台中,以独立的资源共享任务为目标,计算各个任务的时间开销,为每个任务设置优先级,以最坏的任务分配情况下任务的最大响应时间作为共享模型的下限,审核共享资源的安全性以及用户的权限安全性,以云计算平台作为模型的框架,实现政务数据的共享。实验结果表明在链路干扰下和噪声干扰下,设计的共享模型能够保持高效的资源分配性能,资源共享需要的节点映射时间短,仅在20 ms以下,接收请求完整,该共享模型的抗干扰性能得到了提升。

大数据如何影响劳动收入份额?——来自国家级大数据综合试验区的证据

推动劳动收入份额的增长是实现共同富裕的重要保障。文章立足于数据已经成为中国经济发展最重要的生产要素之一这一基本事实,以大数据作为研究切入,基于2011—2020年中国A股上市公司数据,利用国家级大数据综合试验区建立这一外生事件,采用多期双重差分方法系统考察了大数据与劳动收入份额的关系。研究结果显示,大数据能够显著提升企业劳动收入份额,且在经过平行趋势检验、安慰剂检验等一系列稳健性检验后,该结论依然成立。机制分析表明,大数据可以通过直接效应和间接效应影响企业劳动收入份额。其中,直接效应表现为新任务创造和自动化扩张效应,且新任务创造效应有助于企业劳动收入份额的增加,而自动化扩张效应会导致企业劳动收入份额降低。间接效应表现为自动化加深和融资约束缓解效应,且两种间接效应均能有效提升企业劳动收入份额。进一步分析发现,大数据的劳动收入份额提升效应仅存在于东部地区和高市场化水平地区的企业、资本密集型企业、非国有企业以及小规模企业。此外,大数据主要通过缓解企业间的工资收入差距,而非企业内部的工资收入差距来改善劳动收入内部分配结构。文章的研究有助于理解数据要素如何改变要素收入分配结构,为政府部门强化大数据在改善收入差距的作用提供了决策参考。

一种自适应多任务学习方法

传统多任务学习在网络模型和任务权重方面灵活性偏低,不足以应对复杂场景。为此,提出一种自适应网络模型和任务权重的多任务学习方法。首先,运用决策方法学习多任务网络模型结构;其次,利用多任务损失函数动态调整任务权重;最后,在学习到的网络模型和任务权重下训练网络权重,达到特征共享和任务权重之间的权衡。实验结果表明,相比于固定的网络模型和任务权重,提出方法提高了语义分割和表面法线预测性能。

基于混合共享机制的多任务深度学习方法

针对多任务学习的特征提取和任务区分难题,提出基于混合共享机制的多任务深度学习方法。给出硬参数共享网络混合训练方法和依据灵敏性分析的任务相似度分组策略;对组内、组间任务分别应用硬、软参数共享,给出混合共享网络及其相应训练方法;通过MNIST数据集上的实例研究与分析验证该方法的有效性。该方法充分发挥了硬、软参数共享机制的优点,较好刻画了任务的共享与私有特征,提升了多任务学习的性能。

面向专业研究生的数字图像处理与计算机视觉课程教学改革探讨

文章针对面向专业研究生的数字图像处理与计算机视觉课程教学存在的问题,提出了一种多元教学融合法,即以案例教学法展开理论知识教学、以任务驱动教学法来布置课外科研竞赛项目作业、以分享讨论法开展学术交流讲座,制定了偏重实践的全过程性课程考核评价机制,从而激发学生的学习热情和科研创新潜力,增强学生解决复杂实际科研问题的能力,促进教学质量的提升。