基于时序数据的模式发现算法研究

数据库中的知识发现是人工智能领域的一个重要课题,该文针对时序数据中复杂模式的问题,提出了一种新的时序序列模式的逻辑表示法,并设计出一种新的时序序列建模算法。     

Multi-modal diagnosis combining case-based and model-based reasoning: a formal and experimental analysis

Integrating different reasoning modes in the construction of an intelligent system is one of the most interesting and challenging aspects of modern AI. Exploiting the complementarity and the synergy of different approaches is one of the main motivations that led several researchers to investigate the possibilities of building multi-modal reasoning systems, where different reasoning modalities and different knowledge representation formalisms are integrated and combined. Case-Based Reasoning (CBR) is often considered a fundamental modality in several multi-modal reasoning systems; CBR integration has been shown very useful and practical in several domains and tasks. The right way of devising a CBR integration is however very complex and a principled way of combining different modalities is needed to gain the maximum effectiveness and efficiency for a particular task. In this paper we present results (both theoretical and experimental) concerning architectures integrating CBR and Model-Based Reasoning (MBR) in the context of diagnostic problem solving. We first show that both the MBR and CBR approaches to diagnosis may suffer from computational intractability, and therefore a careful combination of the two approaches may be useful to reduce the computational cost in the average case. The most important contribution of the paper is the analysis of the different facets that may influence the entire performance of a multi-modal reasoning system, namely computational complexity, system competence in problem solving and the quality of the sets of produced solutions. We show that an opportunistic and flexible architecture able to estimate the right cooperation among modalities can exhibit a satisfactory behavior with respect to every performance aspect. An analysis of different ways of integrating CBR is performed both at the experimental and at the analytical level. On the analytical side, a cost model and a competence model able to analyze a multi-modal architecture through the analysis of its individual components are introduced and discussed. On the experimental side, a very detailed set of experiments has been carried out, showing that a flexible and opportunistic integration can provide significant advantages in the use of a multi-modal architecture.     

A framework for the analysis of dynamic processes based on Bayesian networks and case-based reasoning

Bayesian networks are knowledge representation schemes that can capture probabilistic relationships among variables and perform probabilistic inference. Arrival of new evidence propagates through the network until all variables are updated. At the end of propagation, the network becomes a static snapshot representing the state of the domain for that particular time. This weakness in capturing temporal semantics has limited the use of Bayesian networks to domains in which time dependency is not a critical factor. This paper describes a framework that combines Bayesian networks and case-based reasoning to create a knowledge representation scheme capable of dealing with time-varying processes. Static Bayesian network topologies are learned from previously available raw data and from sets of constraints describing significant events. These constraints are defined as sets of variables assuming significant values. As new data are gathered, dynamic changes to the topology of a Bayesian network are assimilated using techniques that combine single-value decomposition and minimum distance length. The new topologies are capable of forecasting the occurrences of significant events given specific conditions and monitoring changes over time. Since environment problems are good examples of temporal variations, the problem of forecasting ozone levels in Mexico City was used to test this framework.     

Temporal Bayesian Knowledge Bases – Reasoning about uncertainty with temporal constraints

Time is ubiquitous. Accounting for time and its interaction with change is crucial to modeling the dynamic world, especially in domains whose study of data is sensitive to time such as in medical diagnosis, financial investment, and natural language processing, to name a few. We present a framework that incorporates both uncertainty and time in its reasoning scheme. It is based on an existing knowledge representation called Bayesian Knowledge Bases. It provides a graphical representation of knowledge, time and uncertainty, and enables probabilistic and temporal inferencing. The reasoning scheme is probabilistically sound and the fusion of temporal fragments is well defined. We will discuss some properties of this framework and introduce algorithms to ensure groundedness during the construction of the model. The framework has been applied to both artificial and real world scenarios.     

Building problem-solving environments with the Arches framework

The computational problems that scientists face are rapidly escalating in size and scope. Moreover, the computer systems used to solve these problems are becoming significantly more complex than the familiar, well-understood sequential model on their desktops. While it is possible to re-train scientists to use emerging high-performance computing (HPC) models, it is much more effective to provide them with a higher-level programming environment that has been specialized to their particular domain. By fostering interaction between HPC specialists and the domain scientists, problem-solving environments (PSEs) provide a collaborative environment. A PSE environment allows scientists to focus on expressing their computational problem while the PSE and associated tools support mapping that domain-specific problem to a high-performance computing system.This article describes Arches, an object-oriented framework for building domain-specific PSEs. The framework was designed to support a wide range of problem domains and to be extensible to support very different high-performance computing targets. To demonstrate this flexibility, two PSEs have been developed from the Arches framework to solve problem in two different domains and target very different computing platforms. The Coven PSE supports parallel applications that require large-scale parallelism found in cost-effective Beowulf clusters. In contrast, RCADE targets FPGA-based reconfigurable computing and was originally designed to aid NASA Earth scientists studying satellite instrument data.     

基于动态约束满足框架的强表达时态规划算法

智能规划已经成为人工智能领域最热门的研究主题之一。近年来,智能规划在现实领域的应用越来越广泛,这对规划器的处理能力和效率提出了很大的挑战。以一类强表达时态规划——基于约束区间规划为研究对象,基于动态约束满足框架设计和实现了一个基于约束区间的规划算法LP-TPOP;对算法的可靠性和完备性进行了证明;最后以一个规划实例演示了算法的运行过程。     

Task Planning For A Mobile Robot In An Indoor Environment Using Object-oriented Domain Information

For a mobile robot to be practical, it needs to navigate in dynamically changing environments and manipulate objects in the environment with operating ease. The main challenges to satisfying these requirements in mobile robot research include the collection of robot environment information, storage and organization of this information, and fast task planning based on available information. Conventional approaches to these problems are far from satisfactory due to their requirement of high computation time. In this paper, we specifically address the problems of storage and organization of the environment information and fast task planning in the area of robotic research. We propose an special object-oriented data model (OODM) for information storage and management in order to solve the first problem. This model explicitly represents domain knowledge and abstracts a global perspective about the robot's dynamically changing environment. To solve the second problem, we introduce a fast task planning algorithm that fully uses domain knowledge related to robot applications and to the given environment. Our OODM based task planning method presents a general frame work and representation, into which domain specific information, domain decomposition methods and specific path planners can be tailored for different task planning problems. This method unifies and integrates the salient features from various areas such as database, artificial intelligence, and robot path planning, thus increasing the planning speed significantly     

Post-design analysis for building and refining AI planning systems

The growth of industrial applications of artificial intelligence has raised the need for design tools to aid in the conception and implementation of such complex systems. The design of automated planning systems faces several engineering challenges including the proper modeling of the domain knowledge: the creation of a model that represents the problem to be solved, the world that surrounds the system, and the ways the system can interact with and change the world in order to solve the problem. Knowledge modeling in AI planning is a hard task that involves acquiring the system requirements and making design decisions that can determine the behavior and performance of the resulting system. In this paper we investigate how knowledge acquired during a post-design phase of modeling can be used to improve the prospective model. A post-design framework is introduced which combines a knowledge engineering tool and a virtual prototyping environment for the analysis and simulation of plans. This framework demonstrates that post-design analysis supports the discovery of missing requirements and can guide the model refinement cycle. We present three case studies using benchmark domains and eight state-of-the-art planners. Our results demonstrate that significant improvements in plan quality and an increase in planning speed of up to three orders of magnitude can be achieved through a careful post-design process. We argue that such a process is critical for the deployment of AI planning technology in real-world engineering applications.     

带有双判别器的对抗性领域适应图像分类算法

生成对抗网络的出现将对抗学习的思想引入了机器学习的不同知识体系,带来了全新的发展。对抗性的领域适应算法利用一个共享特征提取器提取域不变表征,一个判别器进行辨别,双方通过对抗性的迭代更新方式达到最优解。在数据来源上,生成对抗网络和领域适应都有极其类似的2个域。在目标函数上,两者都试图追寻一致性。从理论和逻辑结构出发分析两者的内在相似性,尝试利用已成熟的生成对抗网络体系从更深层次进一步提升领域适应性能。通过类比,提出使用2个判别器解决已有对抗性领域适应算法中存在的“模式崩溃”问题,并使用伪标签进行结构上的完善。最后,在标准领域适应任务上的实验表明了本文算法的可行性和有效性。     

Time series representation and similarity based on local autopatterns

Time series data mining has received much greater interest along with the increase in temporal data sets from different domains such as medicine, finance, multimedia, etc. Representations are important to reduce dimensionality and generate useful similarity measures. High-level representations such as Fourier transforms, wavelets, piecewise polynomial models, etc., were considered previously. Recently, autoregressive kernels were introduced to reflect the similarity of the time series. We introduce a novel approach to model the dependency structure in time series that generalizes the concept of autoregression to local autopatterns. Our approach generates a pattern-based representation along with a similarity measure called learned pattern similarity (LPS). A tree-based ensemble-learning strategy that is fast and insensitive to parameter settings is the basis for the approach. Then, a robust similarity measure based on the learned patterns is presented. This unsupervised approach to represent and measure the similarity between time series generally applies to a number of data mining tasks (e.g., clustering, anomaly detection, classification). Furthermore, an embedded learning of the representation avoids pre-defined features and an extraction step which is common in some feature-based approaches. The method generalizes in a straightforward manner to multivariate time series. The effectiveness of LPS is evaluated on time series classification problems from various domains. We compare LPS to eleven well-known similarity measures. Our experimental results show that LPS provides fast and competitive results on benchmark datasets from several domains. Furthermore, LPS provides a research direction and template approach that breaks from the linear dependency models to potentially foster other promising nonlinear approaches.     

Deep continual transfer learning with dynamic weight aggregation for fault diagnosis of industrial streaming data under varying working conditions

Catastrophic forgetting of learned knowledges and distribution discrepancy of different data are two key problems within fault diagnosis fields of rotating machinery. However, existing intelligent fault diagnosis methods generally tackle either the catastrophic forgetting problem or the domain adaptation problem. In complex industrial environments, both the catastrophic forgetting problem and the domain adaptation problem will occur simultaneously, which is termed as continual transfer problem. Therefore, it is necessary to investigate a more practical and challenging task where the number of fault categories are constantly increasing with industrial streaming data under varying operation conditions. To address the continual transfer problem, a novel framework named deep continual transfer learning network with dynamic weight aggregation (DCTLN-DWA) is proposed in this study. The DWA module is used to retain the diagnostic knowledge learned from previous phases and learn new knowledge from the new samples. The adversarial training strategy is applied to eliminate the data distribution discrepancy between source and target domains. The effectiveness of the proposed framework is investigated on an automobile transmission dataset. The experimental results demonstrate that the proposed framework can effectively handle the industrial streaming data under different working conditions and can be utilized as a promising tool for solving actual industrial problem.     

基于P2P覆盖网络的服务领域部署策略

针对服务部署策略不完善的问题,提出P2P覆盖网络框架下自上而下的2级服务部署策略。在顶层P2P网络中部署领域,并在每个领域中部署各种服务组件。对单个领域和伙伴关系领域的放置情况进行建模,使用3种算法进行求解。仿真实验结果表明,伙伴关系领域的邻近放置策略降低了跨领域组合服务的路由开销。     

IPSS: A Hybrid Approach to Planning and Scheduling Integration

Recently, the areas of planning and scheduling in artificial intelligence (AI) have witnessed a big push toward their integration in order to solve complex problems. These problems require both reasoning on which actions are to be performed as well as their precedence constraints (planning) and the reasoning with respect to temporal constraints (e.g., duration, precedence, and deadline); those actions should satisfy the resources they use (scheduling). This paper describes IPSS (integrated planning and scheduling system), a domain independent solver that integrates an AI planner that synthesizes courses of actions with constraint-based techniques that reason based upon time and resources. IPSS is able to manage not only simple precedence constraints, but also more complex temporal requirements (as the Allen primitives) and multicapacity resource usage/consumption. The solver is evaluated against a set of problems characterized by the use of multiple agents (or multiple resources) that have to perform tasks with some temporal restrictions in the order of the tasks or some constraints in the availability of the resources. Experiments show how the integrated reasoning approach improves plan parallelism and gains better makespans than some state-of-the-art planners where multiple agents are represented as additional fluents in the problem operators. It also shows that IPSS is suitable for solving real domains (i.e., workflow problems) because it is able to impose temporal windows on the goals or set a maximum makespan, features that most of the planners do not yet incorporate     

NatureTime: Temporal Granularity in Simulation of Ecosystems

Granularity of time is an important issue for the understanding of how actions performed at coarse levels of time interact with others, working at finer levels. However, it has not received much attention from most AI work on temporal logic. In simpler domains of application we may not need to consider it a problem but it becomes important in more complex domains, such as ecological modelling. In this domain, aggregation of processes working at different time granularities (and sometimes cyclically) is very difficult to achieve reliably. We have proposed a new time granularity theory based onmodular temporal classes, and have developed a temporal reasoning system to specify cyclical processes of simulation models in ecology at many levels of time.     

Probabilistic temporal reasoning and its application to fossil power plant operation

Many real-world applications, such as industrial diagnosis, require an adequate representation and inference mechanism that combines uncertainty and time. In this work, we propose a novel approach for representing dynamic domains under uncertainty based on a probabilistic framework, called temporal nodes Bayesian networks (TNBN). The TNBN model is an extension of a standard Bayesian network, in which each temporal node represents an event or state change of a variable and the arcs represent causal–temporal relationships between nodes. A temporal node has associated a probability distribution for its time of occurrence, where time is discretized in a finite number of temporal intervals; allowing a different number of intervals for each node and a different duration for the intervals within a node (multiple granularity). The main difference with previous probabilistic temporal models is that the representation is based on state changes at different times instead of state values at different times. Given this model, we can reason about the probability of occurrence of certain events, for diagnosis or prediction, using standard probability propagation techniques developed for Bayesian networks. The proposed approach is applied to fossil power plant diagnosis through two detailed case studies: power load increment and control level system failure. The results show that the proposed formalism could help to improve power plant availability through early diagnosis of events and disturbances.     

Temporal Analysis of Motion in Video Sequences through Predictive Operators

We study the problem of recovering temporal parameters which act as predictive operators, generalize time-to-collision and have direct interpretation for navigational purposes for piecewise arbitrarily smooth (polynomial) motion. A result stating that, for monocular observers undergoing arbitrary polynomial laws, these parameters are visually observable, is presented in the first part of this paper. This property suggests an alternate temporal representation of visual looming information. The second part of this paper is concerned with algorithmic approaches for environments with maneuvering agents. A method addressing model order determination, collision detection, and temporal parameter estimation is proposed. Experimental results are reported.     

融合多层注意力的方面级情感分析模型

方面情感分析旨在分析给定文本中特定方面的情感极性。针对目前的研究方法存在对方面情感注意力引入不足问题,提出了一种融合BERT和多层注意力的方面级情感分类模型（BERT and Multi-Layer Attention,BMLA）。模型首先提取BERT内部多层方面情感注意力信息,将编码后的方面信息与BERT隐藏层表征向量融合设计了多层方面注意力,然后将多层方面注意力与编码输出文本进行级联,进而增强了句子与方面词之间的长依赖关系。在SemEval2014 Task4和AI Challenger 2018数据集上的实验表明,强化目标方面权重并在上下文进行交互对方面情感分类是有效的。     

A framework for meta-level control in multi-agent systems

Sophisticated agents operating in open environments must make decisions that efficiently trade off the use of their limited resources between dynamic deliberative actions and domain actions. This is the meta-level control problem for agents operating in resource-bounded multi-agent environments. Control activities involve decisions on when to invoke and the amount to effort to put into scheduling and coordination of domain activities. The focus of this paper is how to make effective meta-level control decisions. We show that meta-level control with bounded computational overhead allows complex agents to solve problems more efficiently than current approaches in dynamic open multi-agent environments. The meta-level control approach that we present is based on the decision-theoretic use of an abstract representation of the agent state. This abstraction concisely captures critical information necessary for decision making while bounding the cost of meta-level control and is appropriate for use in automatically learning the meta-level control policies.     

Continuous frame motion sensitive self-supervised collaborative network for video representation learning

Motion, as a feature of video that changes in temporal sequences, is crucial to visual understanding. The powerful video representation and extraction models are typically able to focus attention on motion features in challenging dynamic environments to complete more complex video understanding tasks. However, previous approaches discriminate mainly based on similar features in the spatial or temporal domain, ignoring the interdependence of consecutive video frames. In this paper, we propose the motion sensitive self-supervised collaborative network, a video representation learning framework that exploits a pretext task to assist feature comparison and strengthen the spatiotemporal discrimination power of the model. Specifically, we first propose the motion-aware module, which extracts consecutive motion features from the spatial regions by frame difference. The global–local contrastive module is then introduced, with context and enhanced video snippets being defined as appropriate positive samples for a broader feature similarity comparison. Finally, we introduce the snippet operation prediction module, which further assists contrastive learning to obtain more reliable global semantics by sensing changes in continuous frame features. Experimental results demonstrate that our work can effectively extract robust motion features and achieve competitive performance compared with other state-of-the-art self-supervised methods on downstream action recognition and video retrieval tasks.     

Utilizing Event-B for domain engineering: a critical analysis

This paper presents our experience of modeling land transportation domain in the formal framework of Event-B. Well-specified requirements are crucial for good software design; they depend on the understanding of the domain. Thus, domain engineering becomes an essential activity. The possibility to have a formal model of a domain, consistent with the use of formal methods for developing critical software working within it, is an important issue. Safety-critical domains, like transportation, exhibit interesting features, such as high levels of nondeterminism, complex interactions, stringent safety properties, and multifaceted timing attributes. The formal representation of these features is a challenging task. We explore the possibility of utilizing Event-B as a domain engineering tool. We discuss the problems we faced during this exercise and how we tackled them. Special attention is devoted to the issue of the validation of the model, in particular with a technique based on the animation of specifications. Event-B is mature enough to be an effective tool to model domains except in some areas, temporal properties mainly, where more work is still needed.