基于条件的边界平衡生成对抗网络

目前没有能够使用简单的网络结构生成高质量特定图像的生成模型,针对这一项任务,本文结合边界平衡生成对抗网络（boundary equilibrium generative adversarial network,BEGAN）的优点,添加附加条件特征以及均方误差损失,建立了条件边界平衡生成对抗网络（conditional-BEGAN,C-BEGAN）,使用这种方法提取其中的生成模型用于特定图像的生成,实验结果表明,该方法相比于其他监督类生成模型可以使用更简单的网络达到更快的收敛速度并且能够生成具有更好质量以及多样性的图片。     

基于CFSM的一致性测试序列生成方法

本文采用通信有限状态机模型描述通信协议,基于通信有限状态机模型提出了协议一致性测试的测试序列生成方法,解决了构件化协议的测试序列生成的问题。本文实现了测试序列的生成算法,通过实例说明了采用测试序列生成算法生成了比传统算法更少的测试序列。同时本算法还可以用于多层协议测试。     

On the Use of Functional Test Generation in Diagnostic Test Generation for Synchronous Sequential Circuits

We study the relationship between diagnostic test generation for a gate-level fault model, which is used for generating diagnostic test sets for manufacturing defects, and functional test generation for a high-level fault model. In general, a functional fault may partially represent some of the effects of one gate-level fault but not another. Generating a test sequence for the functional fault is then likely to detect one gate-level fault but not the other, thus distinguishing the two faults. This relationship points to the ability to use a functional test generation procedure (that targets functional fault detection) as a way of generating diagnostic test sequences for gate-level faults. We use this observation in two ways. The more direct way is to define functional faults that correspond to the differences between pairs of gate-level faults. The second way is to use functional test sequences as diagnostic test sequences without explicitly considering gate-level faults. We support the use of the resulting procedures with experimental results.     

Automatic generation of MPEG test streams from high-level grammars

This paper describes a technique to generate complex, moving picture experts group (MPEG) data streams containing packets which range through a selected set of variants, as allowed by the grammar of the packet stream. The Prolog logic programming language has been used, whose declarative power allows data generation almost directly from the grammar, i.e. without the need for explicitly programming a grammar traversal mechanism as would be the case with an imperative language. A reasonably declarative style of grammar and variation definition is achieved, and at the same time, a reasonably efficient generation process. The basic idea is to use a declarative fragment of Prolog for the grammar, but to use imperative features of Prolog for matters like packet enumeration and packet payload generation. Generation of test data from grammars is not new, nor is the use of Prolog programs for generation of test data, but as far as we know, the combination of both has not reported on in the literature, nor its application to MPEG demultiplexers/decoders.     

Operations Research Methods for Autonomic Resource Management

Varying service demands on next generation networks requires autonomic resource management capabilities to ensure service delivery to customers, motivating the need to develop self-managing algorithms for this purpose. In this paper, we use algorithms from the Operations Research community for this purpose. First, we adapt the Transportation Model to manage distributed resources in the next generation network infrastructure. Coupled with monitoring capabilities, the proposed scheme can automatically adjust the virtual resource allocation to optimize the costs incurred by service providers in offering services to customers. Second, we propose the use of Inventory Control to predict needs for virtual resources and to pre-order required virtual resource amounts. We compare two possible Inventory Control models to manage the virtual resources involved in service delivery. We perform extensive simulations to show the performance improvements made possible by the use of the Transportation Model and Inventory Control for autonomic resource management in next generation networks.     

The Monte-Carlo generation of random polygons

Random surface-covering aggregates of polygons are of interest as a stochastic model for several geological applications. The approach and algorithms for the generation of random polygons are presented along with a FORTRAN program designed to generate the random Voronoi polygons. To demonstrate the use of the program, the results of the generation of 57,000 Voronoi are tabulated in histogram form, for use in hypothesis testing. The computer program can be adapted easily to other random polygon generations.     

Net generation or Digital Natives: Is there a distinct new generation entering university?

This article reports key findings from the first phase of a research project investigating Net generation age students as they encounter e-learning at five universities in England. We take a critical view of the idea of a distinct generation which has been described using various terms including Net generation and Digital Natives and explore age related differences amongst first year university students. The article draws on evidence from a survey of first year undergraduates studying a range of pure and applied subjects. Overall we found a complex picture amongst first-year students with the sample population appearing to be a collection of minorities. These included a small minority that made little use of some technologies and larger minorities that made extensive use of new technologies. Often the use of new technology was in ways that did not fully correspond with the expectations that arise from the Net generation and Digital Natives theses. The article concludes that whilst there are strong age related variations amongst the sample it is far to simplistic to describe young first-year students born after 1983 as a single generation. The authors find that the generation is not homogenous in its use and appreciation of new technologies and that there are significant variations amongst students that lie within the Net generation age band.     

面向视觉语言理解与生成的多模态预训练方法

大多数现有的视觉语言预训练方法侧重于理解任务,并在训练时使用类似于BERT的损失函数(掩码语言建模和图像文本匹配).尽管它们在许多理解类型的下游任务中表现良好,例如视觉问答、图像文本检索和视觉蕴涵,但它们不具备生成信息的能力.为了解决这个问题,提出了视觉语言理解和生成的统一多模态预训练(unified multimodal pre-training for vision-language understanding and generation, UniVL). UniVL能够处理理解任务和生成任务,并扩展了现有的预训练范式,同时使用随机掩码和因果掩码,因果掩码即掩盖未来标记的三角形掩码,这样预训练的模型可以具有自回归生成的能力.将几种视觉语言理解任务规范为文本生成任务,并使用基于模版提示的方法对不同的下游任务进行微调.实验表明,在使用同一个模型时,理解任务和生成任务之间存在权衡,而提升这两个任务的可行方法是使用更多的数据. UniVL框架在理解任务和生成任务方面的性能与最近的视觉语言预训练方法相当.此外,实验还证明了基于模版提示的生成方法更有效,甚至在少数场景中它优于判别方法.     

Multimodal Pre-training Method for Vision-language Understanding and Generation

Most existing vision-language pre-training methods focus on understanding tasks and use BERT-like loss functions (masked language modeling and image-text matching) during pre-training. Despite their good performance in the understanding of downstream tasks, such as visual question answering, image-text retrieval, and visual entailment, these methods cannot generate information. To tackle this problem, this study proposes Unified multimodal pre-training for Vision-Language understanding and generation (UniVL). The proposed UniVL is capable of handling both understanding tasks and generation tasks. It expands existing pre-training paradigms and uses random masks and causal masks simultaneously, where causal masks are triangular masks that mask future tokens, and such pre-trained models can have autoregressive generation abilities. Moreover, several vision-language understanding tasks are turned into text generation tasks according to specifications, and the prompt-based method is employed for fine-tuning of different downstream tasks. The experiments show that there is a trade-off between understanding tasks and generation tasks when the same model is used, and a feasible way to improve both tasks is to use more data. The proposed UniVL framework attains comparable performance to recent vision-language pre-training methods in both understanding tasks and generation tasks. Moreover, the prompt-based generation method is more effective and even outperforms discriminative methods in few-shot scenarios.     

Two case studies in grammar-based test generation

Grammar-based test generation (GBTG) has seen extensive study and practical use since the 1970s. GBTG was introduced to generate source code for testing compilers from context-free grammars specifying language syntax. More recently, GBTG has been applied to many other testing problems, including the generation of eXtensible Markup Language (XML) documents and the generation of packets for testing communications protocols. Recent research has shown how to integrate covering-array techniques such as pairwise testing into GBTG tools. While the integration offers considerable power to the tester, there are few practical demonstrations in the literature. We present two case studies showing how to use grammars and covering arrays for automated software testing. The first case study exposes HTML injection vulnerabilities in an RSS feed parser. The second case study determines the effectiveness of network firewalls when faced with TCP flag attacks. The case studies illustrate the use of covering arrays in a GBTG context, the use of visualization to understand large test logs, and the issues and tradeoffs in the design of fully automated GBTG test suites.     

Generation and synchronous tree-adjoining grammars

Tree-adjoining grammars (TAG) have been proposed as a formalism for generation based on the intuition that the extended domain of syntactic locality that TAGs provide should aid in localizing semantic dependencies as well, in turn serving as an aid to generation from semantic representations. We demonstrate that this intuition can be made concrete by using the formalism of synchronous tree-adjoining grammars. The use of synchronous TAGs for generation provides solutions to several problems with previous approaches to TAG generation. Furthermore, the semantic monotonicity requirement previously advocated for generation grammars as a computational aid is seen to be an inherent property of synchronous TAGs.     

Nationwide Freight Generation Models: A Spatial Regression Approach

This paper investigates the application of linear regression models and modeling techniques in predicting freight generation at the national level within the U.S. Specifically, the paper seeks to improve the performance and fit of linear regression models of freight generation. We provide insight into different variable transformation techniques, evaluate the use of spatial regression variables, and apply a spatial regression modeling methodology to correct for spatial autocorrelation. We conclude that the spatial regression model is the preferred specification for freight generation at the national level. The proliferation of Geographic Information Systems (GIS) within planning agencies affords more widespread use of spatial regression and our results indicate this technique would provide improvement to models that have been traditionally limited by insufficient data.     

Computer generation of sensitivity functions for sampled-data control systems†

Methods have been presented previously (Suryanarayanan and Soudack 1970) for the generation of sensitivity functions for general non-linear sampled-data systems. In this paper, it is shown that economy of simulator components (for hybrid computer generation) or of computational time (for digital computer generation) is achieved for simultaneously generating a number of sensitivity functions in a class of linear sampled-data control systems. Extensive use of signal flow graph algebra is made for this purpose and comprehensive examples are presented.     

测试数据自动生成方法

软件测试是提高软件可靠性、保证软件质量的重要手段,可分为静态分析、路径选择、测试数据生成和动态分析四个阶段,而软件测试过程中的一个重要任务是生成测试数据。文中首先给出了遗传算法的形式化描述,然后提出了遗传算法和函数极小化相结合的方法自动生成测试数据,并通过具体实例表明其有效性。     

遗传算法在网络在线智能组卷中的应用研究

速度和质量是网络在线实时组卷的两个核心要求,常用二进制编码遗传算法组卷时间受题库试题总量影响大,且后期收敛效率低;通过对上述问题的分析．提出了一种基于分段实数代号编玛和微量变异算子的GAs组卷算法．详细描述了新算法的设计思想和实现过程;结果验征了该算法运用于网培在线实时快速组卷的可行性,新算法的收敛时间（组卷时间）不受题库题量影响,可短至1.56sa。     

Web应用前后端融合的遗传算法并行化测试用例生成

Web应用测试用例生成并行化是提升Web应用测试生成效率的一个有效手段.Web应用的前后端分离、事件驱动等特性,导致传统的并行化技术难以直接应用于Web应用的测试用例自动生成中.因此,如何针对Web应用进行并行化测试用例生成,是一项具有挑战性的工作.将种群并行化计算引入到基于遗传算法的Web应用前后端融合的测试用例生成中,通过线程池及调度逻辑设计、多浏览器进程管理及后端覆盖路径获取,实现种群个体在多浏览器上的并行化执行及基于后端路径覆盖的适应度值并行化计算,以更高效地生成Web应用的测试用例.实验结果表明:相对于Web应用的GA串行化测试用例生成方法,所提的并行化测试生成方法能够更充分地利用系统资源,极大地提升Web应用测试用例的生成效率.     

分布式状态空间生成的设计与实现

状态空间生成的并行化是针对状态空间爆炸问题而提出的一种重要手段。提出了一种基于MapReduce的分布式状态空间生成方案,与现有的同类研究相比,它无需用户关心生成算法的并行化,具有简单易用性;与常规的MapReduce的用法相比,它增加了输入文件的自动生成和作业运行的自动循环控制。该方案已在小规模分布式环境下实现,实验结果表明：（1）基于Map-Reduce的分布式状态空间生成算法可以扩大模型的可求解规模;（2）对于状态空间规模的增长主要由托肯（token）数增加引起的一类模型,该算法具有良好的适应性和可扩展性。     

基于声学统计建模的语音合成技术研究

该文介绍基于声学统计建模的语音合成技术,重点回顾中国科学技术大学讯飞语音实验室在语音合成领域这一前沿发展方向的创新性工作成果。具体包括 融合发音动作参数与声学参数,提高声学参数生成的灵活性;以最小生成误差准则取代最大似然准则,提高合成语音的音质;使用单元挑选与波形拼接方法取代参数合成器重构,改善参数语音合成器在合成语音音质上的不足。以上技术创新使得语音合成系统在自然度、表现力、灵活性及多语种应用等方面的性能都有进一步的提升,并推动语音合成技术在呼叫中心信息服务、移动嵌入式设备人机语音交互、智能语音教学等领域的广泛引用。     

A framework for variable content document generation with multiple actors

ContextAdvances in customization have highlighted the need for tools supporting variable content document management and generation in many domains. Current tools allow the generation of highly customized documents that are variable in both content and layout. However, most frameworks are technology-oriented, and their use requires advanced skills in implementation-related tools, which means their use by end users (i.e. document designers) is severely limited.ObjectiveStarting from past and current trends for customized document authoring, our goal is to provide a document generation alternative in which variants are specified at a high level of abstraction and content reuse can be maximized in high variability scenarios.MethodBased on our experience in Document Engineering, we identified areas in the variable content document management and generation field open to further improvement. We first classified the primary sources of variability in document composition processes and then developed a methodology, which we called DPL – based on Software Product Lines principles – to support document generation in high variability scenarios.ResultsIn order to validate the applicability of our methodology we implemented a tool – DPLfw – to carry out DPL processes. After using this in different scenarios, we compared our proposal with other state-of-the-art tools for variable content document management and generation.ConclusionThe DPLfw showed a good capacity for the automatic generation of variable content documents equal to or in some cases surpassing other currently available approaches. To the best of our knowledge, DPLfw is the only framework that combines variable content and document workflow facilities, easing the generation of variable content documents in which multiple actors play different roles.