ID	原文	译文
56608	因此, 本文探索了意图敏感的日志增强方法, 提出了一种日志意图描述模型 (log-intention description model, LIDM), 在此基础上设计和实现了自动化日志增强工具 SmartLog.SmartLog 利用 LIDM 提取日志代码意图, 挖掘日志增强规则, 进而实现意图敏感的日志自动增强.	To achieve this, we propose a log intention descriptionmodel to describe the intention of log statements. SmartLog then explores the intentions of existing logs andmines log rules from those intentions.
56609	本文在 6 款成熟且被广泛使用的开源软件上对 SmartLog 的有效性展开了评估.	We evaluated SmartLog on six mature open-source projects.
56610	评估结果显示, SmartLog 的准确性相比两个已有最好的日志增强工具分别提升 43% 和 16%.	Comparedwith two state-of-the-art projects, i. e. , Errlog and LogAdvisor, SmartLog improved the accuracy of log placementby 43% and 16% respectively.
56611	此外, 本文收集了软件演化过程中 86 个开发人员增加日志的实例, 并使用 SmartLog 和两个已有工具分析每次日志演化的旧软件版本, 发现 3 个工具可自动在新软件版本添加的日志分别是 49, 10, 22 个, 软件演化效率相比已有工作显著增强.	SmartLog could cover 49 out of 86 real-world patches aimed to add logs, while thestate-of-the-art works could cover 10 and 22 patches, respectively.
56612	随着人工智能和物联网的快速发展与融合,智能物联网AIoT正成长为一个极具前景的新兴前沿领域,其中深度学习模型的终端运行是其主要特征之一.	The rapid development of both Artificial Intelligence (AI) and the Internet of Things (IoT), hascultivated the new research area: the Artificial Intelligence of Things (AIoT). AIoT is used to deploy manydifferent deep learning models on a variety of local IoT terminals including smartphones, wearables, and otherembedded devices.
56613	针对智能物联网应用场景动态多样,以及物联网终端(智能手机、可穿戴及其他嵌入式设备等)计算和存储资源受限等问题,深度学习模型环境自适应正成为一种新的模型演化方式.	Adapting to these dynamic and varied AIoT application scenarios, and the IoT platformresources (e. g. , computation and storage resources) available in each diverse, requires a novel scheme for improvingon device environmental adaptability.
56614	其旨在确保适当性能的条件下,能自适应地根据环境变化动态调整模型,从而降低资源消耗、提高运算效率.	Deep learning models aim to dynamically adjust either the model structure,the calculation scheme, or both, of them specifically to adapt to the environment context. They must reducecosts and improve computational efficiency while creating negligible performance degradation.
56615	具体来说,它需要主动感知环境、任务性能需求和平台资源约束等动态需求,进而通过终端模型的自适应压缩、云边端模型分割、领域自适应等方法,实现深度学习模型对终端环境的动态自适应和持续演化.	Specifically, anenvironmental adaptation evolution framework must actively and continuously assess the constantly changingenvironmental context including factors, such as application data, knowledge base, task-related performancerequirements, and platform-imposed resource constraints. Then it must adopt on-demand model compression,model segmentation, and domain adaptation techniques to achieve a appropriate balance between the model’sperformance and the environment’s budget.
56616	本文围绕深度学习模型自适应问题,从其概念、系统架构、研究挑战与关键技术等不同方面进行阐述和讨论,并介绍我们在这方面的研究实践.	This paper focuses on making deep learning models for context-awareadaptation. We discuss the system architecture and core technologies solving this problem requires. We addressresearch challenges in this area, and introduce our pilot research practice in this field.
56617	围绕如何构建能够在高动态战场环境下持续保障多样化任务完成的指挥控制系统,依据可成长软件理论、方法,提出了一套指挥控制系统的自主适变解决方案;	Command and control systems used in highly dynamic battlefield environments must be built tocontinuously guarantee the completion of diverse missions. This paper presents a novel solution for buildingautonomous adaptive command and control systems, based on the theories, and methods of growing softwaresystems.