Deep learning for procedural content generation

Neural Computing and Applications - Procedural content generation in video games has a long history. Existing procedural content generation methods, such as search-based, solver-based, rule-based...     

A fast approach for automatic generation of fuzzy rules bygeneralized dynamic fuzzy neural networks

A fast approach for automatically generating fuzzy rules from sample patterns using generalized dynamic fuzzy neural networks (GD-FNNs) is presented. The GD-FNN is built based on ellipsoidal basis functions and functionally is equivalent to a Takagi-Sugeno-Kang fuzzy system. The salient characteristics of the GD-FNN are: (1) structure identification and parameters estimation are performed automatically and simultaneously without partitioning input space and selecting initial parameters a priori; (2) fuzzy rules can be recruited or deleted dynamically; (3) fuzzy rules can be generated quickly without resorting to the backpropagation (BP) iteration learning, a common approach adopted by many existing methods. The GD-FNN is employed in a wide range of applications ranging from static function approximation and nonlinear system identification to time-varying drug delivery system and multilink robot control. Simulation results demonstrate that a compact and high-performance fuzzy rule-base can be constructed. Comprehensive comparisons with other latest approaches show that the proposed approach is superior in terms of learning efficiency and performance     

面向系统动态可靠性的自适应目标代码生成方法

随着社会的快速发展,人们对嵌入式智能设备的依赖程度越来越高。嵌入式系统工作在体积受限的封闭环境中,运算部件、存储单元等相关元器件体积小集成度高,不同工作环境、不同使用频度对电子元器件的可靠性将产生重要影响。针对嵌入式系统工作过程中的动态可靠性,提出了面向系统动态可靠性的自适应目标代码生成方法。该方法借助于决策树学习算法,构建了系统可靠性评估模型;并以此为指导,设计了多路径目标代码生成方法。使得系统能够根据实际工作状态信息,自适应的最佳的执行路径,以避免系统资源使用的不均衡,提高各运算部件的可靠性。实验表明,该方法将程序对单个处理器最高使用率由80%以上降到了30%以内,将内存单元最大最小访问比例由157.3降到了15.4,有效均衡了各处理器核和内存单元的使用。     

交通出行诱导的离散Stackelberg动态博弈模型及其求解算法

针对交通出行诱导的实际需要,分析了道路交通管理者和道路使用者的博弈策略和博弈特征.通过道路使用者信息的模糊模型.建立了交通出行诱导的离散动态Stackelberg博弈模型.分析了交通出行诱导的两阶段博弈特征,提出了基于逆向归纳法的博弈模型求解算法.通过对模型无约束转化,运用Monte-Carlo法对示例路网进行了计算机求解,求解结果表明产生的交通出行诱导方案能够实现系统最优下的用户最优.     

基于列生成算法求解动态库存板坯分配问题

针对钢铁企业中遇到的动态库存板坯分配问题进行了研究。建立了一个0-1整数规划数学模型,该模型的目标是最小化板坯与合同规格差异费用以及板坯在库停留所产生的库存成本费用之和。根据问题特点,使用Danzig-Wolfe策略将这个模型分解为一个带有集划分约束的主问题和一个具有背包特征约束的价格子问题,开发了分支价格算法进行求解。计算结果表明所开发的分支价格算法能够最优求解生产实际问题。     

基于动态双极值模糊软集的群决策方法

针对实际问题中双极值模糊软集随时间变化的影响,定义了动态双极值模糊软集等概念,讨论了相关运算及性质。根据时间权重符合对数增长模型得到权重确定公式。利用集成思想定义双极值模糊软集的运算并给出几何加权平均算子的计算公式,将动态双极值模糊软集集成为综合双极值模糊软集。利用水平软集算出各对象的机会值,得出最优决策。通过实例分析证明此决策方法的合理性与可行性。     

An interactive method for dynamic intuitionistic fuzzy multi-attribute group decision making

This paper investigates the dynamic intuitionistic fuzzy multi-attribute group decision making (DIF-MAGDM) problems, in which all the attribute values provided by multiple decision makers (DMs) at different periods take the form of intuitionistic fuzzy numbers (IFNs), and develops an interactive method to solve the DIF-MAGDM problems. The developed method first aggregates the individual intuitionistic fuzzy decision matrices at different periods into an individual collective intuitionistic fuzzy decision matrix for each decision maker by using the dynamic intuitionistic fuzzy weighted averaging (DIFWA) operator, and then employs intuitionistic fuzzy TOPSIS method to calculate the individual relative closeness coefficient of each alternative for each decision maker and obtain the individual ranking of alternatives. After doing so, the method utilizes the hybrid weighted averaging (HWA) operator to aggregate all the individual relative closeness coefficients into the collective relative closeness coefficient of each alternative and obtain the aggregate ranking of alternatives, by which the optimal alternative can be selected. In addition, the spearman correlation coefficient for both the aggregate ranking and individual ranking of alternatives is calculated to measure the consensus level of the group preferences. Finally, a numerical example is used to illustrate the developed method.     

一种实用的动态模糊度表示方法

文章分析了动态模糊逻辑理论中动态模糊数据的构成，根据实际需要，将动态模糊数据分离成模糊数据和动态数据进行表示，提出了一种新的动态度的表示方法，并给出了动态度的近似计算方法。     

A new identification method for fuzzy linear models of nonlinear dynamic systems

The most promising methods for identifying a fuzzy model are data clustering, cluster merging and subsequent projection of the clusters on the input variable space. This article proposes to modify this procedure by adding a cluster rotation step, and a method for the direct calculation of the consequence parameters of the fuzzy linear model. These two additional steps make the model identification procedure more accurate and limits the loss of information during the identification procedure. The proposed method has been tested on a nonlinear first order model and a nonlinear model of a bioreactor and results are very promising.     

基于模糊规则的学习者认知水平估算方法

认知水平估算是远程学习系统中实现学习导航和知识推荐的一个重要依据.针对现有的估算方法存在计算方式单一、逻辑推理简单等问题,提出了一种用模糊集表示认知水平等级,以记忆量为判断标准、试题难度和做题时间为影响因素,通过模糊推理产生对认知等级不同的影响参数,并利用影响参数和改进的SherlockⅡ方法修改认知水平的等级隶属度的方法.通过和其他方法比较,得出本方法逻辑合理、计算简便,能够合理地反映学习者的认知过程.     

Comparison of Yager's level set method for fuzzy logic control withMamdani's and Larsen's methods

A new fuzzy reasoning method for fuzzy control recently proposed by R. Yager is investigated. A comparison with the most useful fuzzy control schemes, for a first-order with time delay process, is carried out. The results obtained show that Yager's method is superior from the point of view of both computational burden and control system behavior     

Automatic evolution of programs for procedural generation of terrains for video games

Nowadays the video game industry is facing a big challenge to keep costs under control as games become bigger and more complex. Creation of game content, such as character models, maps, levels, textures, sound effects and so on, represent a big slice of total game production cost. Hence, the video game industry is increasingly turning to procedural content generation to amplify the cost-effectiveness of the efforts of video game designers. However, procedural methods for automated content generation are difficult to create and parametrize. In this work we study a genetic programming-based procedural content technique to generate procedural terrains that do not require parametrization, thus, allowing to save time and help reducing production costs. Generated procedural terrains present aesthetic appeal; however, unlike most techniques involving aesthetic, our approach does not require a human to perform the evaluation. Instead, the search is guided by the weighted sum of two morphological metrics: terrain accessibility and obstacle edge length. The combination of the two metrics allowed us to find a wide range of fit terrains that present more scattered obstacles in different locations than our previous approach with a single metric. Procedural terrains produced by this technique are already in use in a real video game.     

Knowledge-based fuzzy MLP for classification and rule generation

A new scheme of knowledge-based classification and rule generation using a fuzzy multilayer perceptron (MLP) is proposed. Knowledge collected from a data set is initially encoded among the connection weights in terms of class a priori probabilities. This encoding also includes incorporation of hidden nodes corresponding to both the pattern classes and their complementary regions. The network architecture, in terms of both links and nodes, is then refined during training. Node growing and link pruning are also resorted to. Rules are generated from the trained network using the input, output, and connection weights in order to justify any decision(s) reached. Negative rules corresponding to a pattern not belonging to a class can also be obtained. These are useful for inferencing in ambiguous cases. Results on real life and synthetic data demonstrate that the speed of learning and classification performance of the proposed scheme are better than that obtained with the fuzzy and conventional versions of the MLP (involving no initial knowledge encoding). Both convex and concave decision regions are considered in the process.     

一种军棋机器博弈的多棋子协同博弈方法

针对在军棋博弈不完全信息对弈中,面对棋子不同价值、不同位置、不同搭配所产生的不同棋力,传统的单子意图搜索算法,不能满足棋子之间的协同性与沟通性,同时也缺乏对敌方的引诱与欺骗等高级对抗能力。本文提出一种结合UCT搜索策略的高价值棋子博弈方法,实现高价值棋子协同博弈的策略。实战经验表明:高价值多棋子军棋协同博弈策略优于单棋子军棋博弈策略。     

Coalitional game with fuzzy payoffs and credibilistic core

As an important branch of game theory, coalitional game deals with situations that involve cooperation among the players. This paper deals with this topic further by incorporating the fuzzy payoff information. Based on the credibility theory, we introduce two decision criteria to define the preferences of players, which leads to two definitions of credibilistic cores—the solution of coalitional game with fuzzy transferable payoffs. Meanwhile, we give a sufficient and necessary condition to ensure non-emptiness of the credibilistic cores. Finally, an example is provided for illustrating the usefulness of the theory developed in this paper.     

Automatic fuzzy ontology generation for semantic Web

Ontology is an effective conceptualism commonly used for the semantic Web. Fuzzy logic can be incorporated to ontology to represent uncertainty information. Typically, fuzzy ontology is generated from a predefined concept hierarchy. However, to construct a concept hierarchy for a certain domain can be a difficult and tedious task. To tackle this problem, this paper proposes the FOGA (fuzzy ontology generation framework) for automatic generation of fuzzy ontology on uncertainty information. The FOGA framework comprises the following components: fuzzy formal concept analysis, concept hierarchy generation, and fuzzy ontology generation. We also discuss approximating reasoning for incremental enrichment of the ontology with new upcoming data. Finally, a fuzzy-based technique for integrating other attributes of database to the ontology is proposed.     

基于CNN和BiLSTM的短文本相似度计算方法

文本语义相似度学习在自然语言处理领域扮演着重要角色。传统的基于统计学习的方法只能对句子的浅层语义进行建模,在深层语义上效果表现不佳。针对统计学习方法无法有效解决深层语义的问题,提出一种基于卷积神经网络(Convolutional Neural Networks,CNN)和双向长短期记忆(Bidirectional Long Short Term Memory,BiLSTM)相结合的短文本相似度计算方法,利用CNN与BiLSTM相结合的孪生神经网络结构提取句子不同粒度的特征,对句子进行深层语义编码。此外,结合attention机制将提取到的单词级别的特征组合拼接成句子的语义向量表示,最后通过匹配层来判断文本的语义相似度。实验结果表明,基于Quora Questions Pairs数据集,所提出的短文本相似度计算方法准确率为84. 58%,F1值为85. 02%,优于其他方法。     

Rotational perspective and learning procedural tasks from dynamic media

There have been conflicting accounts regarding the effectiveness of animations for learning. Procedural motor learning represents one of the few areas in which animations have consistently shown to be facilitative. Some have suggested that this benefit is related to activation of the mirror neuron system (MNS), with higher activation leading to better performance. This study examines this explanation, and observed the effects of instructional media (animation vs. static), as a function of viewing perspective (face-to-face vs. over-the-shoulder) on understanding a procedural motor task (knot tying). Results indicate that performance was significantly improved with animations over static images, however this appeared to be most pronounced in situations which matched the learners' own perspective (i.e., over-the-shoulder). These findings have implications for the design of instructional media for procedural motor tasks and provide confirmation of the assertion that appropriate activation of the perceptual system can be leveraged to increase performance.     

直觉模糊语言信息下的三维动态评价方法及应用

针对当前综合评价大多集中在静态评价以及属性权重未知的问题,在直觉模糊语言环境下提出一种三维动态评价方法.首先,利用模糊质量功能展开求解指标权重以弥补传统主客观确权方法的不足,体现指标越能表征顾客需求贡献度权重越大的赋权思路;其次,考虑到评价指标间的相互关系也会影响最终的评价结果,引入幂平均算子并将其扩展到直觉模糊语言集中,通过直觉模糊语言幂加权平均算子集结多属性信息得到静态综合评价值;接着,在静态评价值求解的基础上,基于全局信息定义分层激励因子和阶段信息定义成长因子分别对评价基础值的增长态势和增长趋势进行测度,构建基于评价基础值、变化状态与发展趋势的三维动态评价指数模型;最后以小微企业的信用评价为例验证所提出方法的有效性和可行性.