node_id int64 0 76.9k | label int64 0 39 | text stringlengths 13 124k | neighbors listlengths 0 3.32k | mask stringclasses 4
values |
|---|---|---|---|---|
780 | 0 | Multiagent Systems Specification by UML Statecharts Aiming at Intelligent Manufacturing Multiagent systems are a promising new paradigm in computing, which are contributing to various fields. Many theories and technologies have been developed in order to design and specify multiagent systems, however, no standard procedure is used at present. Industrial applications often have a complex structure and need plenty of working resources. They require a standard specification method as well. As the standard method to design and specify software systems, we believe that one of the key words is simplicity for their wide acceptance. In this paper, we propose a method to specify multiagent systems, namely with UML statecharts. We use them for specifying almost all aspects of multiagent systems, because we think that it is an advantage to keep everything in one type of diagram. We apply | [
403,
1129,
1367,
1583
] | Train |
781 | 3 | A Performance Evaluation of Alternative Mapping Schemes for Storing XML Data in a Relational Database XML is emerging as one of the dominant data formats for data processing on the Internet. To query XML data, query languages likeXQL, Lorel, XML-QL, or XML-GL have been proposed. In this paper, we study how XML data can be stored and queried using a standard relational database system. For this purpose, we present alternative mapping schemes to store XML data in a relational database and discuss how XML-QL queries can be translated into SQL queries for every mapping scheme. We present the results of comprehensive performance experiments that analyze the tradeo#s of the alternative mapping schemes in terms of database size, query performance and update performance. While our discussion is focussed on XML and XML-QL, the results of this paper are relevant for most semi-structured data models and most query languages for semi-structured data. 1 Introduction It has become clear that not all applications are met by the relational, object-relational, or object-oriented data models. ... | [
29,
176,
661,
1453,
3118
] | Train |
782 | 0 | Aspects of Interface Agents: Avatar, Assistant and Actor This paper introduces the interface agent research project being carryied out at ATR Media Integration & Communications Research Labs. We are interested in virtual interface agents to support human creative activities by mediating between human users and computer cyberspace. In this paper, we categorize interface agents into three types; avatars, assistants and actors based on their functionality and discuss a design framework of agents and related applications. We present the current status and objectives for the following topics; design of an asynchronous-hierarchical agent architecture, a character locomotion design tool, applications in a virtualized museum and a group discussion environment in which virtual agents would inhabit and act. 1 Introduction A "virtual interface agent," or interface agent, is defined in this paper as an autonomous agent which mediates between a human user and computer cyberspace. An interface agent differs from an ordinary interface since it is expecte... | [
1216
] | Test |
783 | 2 | Text-Based Content Search and Retrieval in ad hoc P2P Communities We consider the problem of content search and retrieval in peer-to-peer (P2P) communities. P2P computing is a potentially powerful model for information sharing between ad hoc groups' of users because of its' low cost of entry and natural model for resource scaling with community size. As P2P communities grow in size, however, locating information distributed across the large number of peers becomes problematic. We present a distributed text-based content search and retrieval algorithm to address this' problem. Our algorithm is' based on a state-of-the-art text-based document ranking algorithm: the vector-space model instantiated with the TFxlDF ranking rule. A naive application of TFxlDF wouM require each peer in a community to collect an inverted index of the entire community. This' is' costly both in terms of bandwidth and storage. Instea & we show how TFxlDF can be approximated given compact summaries of peers' local inverted indexes. We make three contributions: (a) we show how the TFxlDF rule can be adapted to use the index summaries, (b) we provide a heuristic for adaptively determining the set of peers that shouM be contacted for a query, and (c) we show that our algorithm tracks' TFxlDF's performance very closely, regardless of how documents' are distributed throughout the community. Furthermore, our algorithm preserves the main flavor of TFxlDF by retrieving close to the same set of documents for any given query. | [
1505,
2503
] | Validation |
784 | 5 | Improving Data Driven Wordclass Tagging by System Combination In this paper we examine how the differences in modelling between different data driven systems performing the same NLP task can be exploited to yield a higher accuracy than the best individual system. We do this by means of an experiment involving the task of morpho-syntactic wordclass tagging. Four well-known tagger generators (Hidden Markov Model, Memory-Based, Transformation Rules and Maximum Entropy) are trained on the same corpus data. After comparison, their outputs are combined using several voting strategies and second stage classifiers. All combination taggers outperform their best component, with the best combination showing a 19.1% lower error rate than the best individual tagger. Introduction In all Natural Language Processing (NLP) systems, we find one or more language models which are used to predict, classify and/or interpret language related observations. Traditionally, these models were categorized as either rule-based/symbolic or corpusbased /probabilistic. Recent ... | [
2676
] | Train |
785 | 1 | Saccadic Search with Gabor features applied to Eye Detection and Real-Time Head Tracking The Gabor decomposition is a ubiquitous tool in computer vision. Nevertheless, it is generally considered computationally demanding for active vision applications. We suggest an attention-driven approach to feature detection inspired by the human saccadic system. A dramatic speedup is achieved by computing the Gabor decomposition only on the points of a sparse retinotopic grid. An off-line eye detection application and a real-time head localisation and tracking system are presented. The real-time system features a novel eyeball-mounted camera designed to simulate the dynamic performance of the human eye and is, to the best of our knowledge, the first example of active vision system based on the Gabor decomposition. | [
2887
] | Train |
786 | 1 | A Mutually Beneficial Integration of Data Mining and Information Extraction Text mining concerns applying data mining techniques to unstructured text. Information extraction (IE) is a form of shallow text understanding that locates specific pieces of data in natural language documents, transforming unstructured text into a structured database. This paper describes a system called DISCOTEX, that combines IE and data mining methodologies to perform text mining as well as improve the performance of the underlying extraction system. Rules mined from a database extracted from a corpus of texts are used to predict additional information to extract from future documents, thereby improving the recall of IE. Encouraging results are presented on applying these techniques to a corpus of computer job announcement postings from an Internet newsgroup. | [
327,
1569,
2068,
2750
] | Train |
787 | 4 | How Many Separately Evolved Emotional Beasties Live Within Us? A problem which bedevils the study of emotions, and the study of consciousness, is that we assume a shared understanding of many everyday concepts, such as `emotion', `feeling', `pleasure', `pain', `desire', `awareness', etc. Unfortunately, these concepts are inherently very complex, ill-defined, and used with different meanings by different people. Moreover this goes unnoticed, so that people think they understand what they are referring to even when their understanding is very unclear. Consequently there is much discussion that is inherently vague, often at cross-purposes, and with apparent disagreements that arise out of people unwittingly talking about different things. We need a framework which explains how there can be all the diverse phenomena that different people refer to when they talk about emotions and other affective states and processes. The conjecture on which this paper is based is that adult humans have a type of information-processing architecture, with components whi... | [
3143
] | Train |
788 | 2 | Summarizing Text Documents: Sentence Selection and Evaluation Metrics Human-quality text summarization systems are difficult to design, and even more difficult to evaluate, in part because documents can differ along several dimensions, such as length, writing style and lexical usage. Nevertheless, certain cues can often help suggest the selection of sentences for inclusion in a summary. This paper presents our analysis of news-article summaries generated by sentence selection. Sentences are ranked for potential inclusion in the summary using a weighted combination of statistical and linguistic features. The statistical features were adapted from standard IR methods. The potential linguistic ones were derived from an analysis of news-wire summaries. Toevaluate these features we use a normalized version of precision-recall curves, with a baseline of random sentence selection, as well as analyze the properties of such a baseline. We illustrate our discussions with empirical results showing the importance of corpus-dependent baseline summarization standards, compression ratios and carefully crafted long queries. | [
1152,
1241,
1640,
1753
] | Validation |
789 | 0 | Standardizing Agent Communication An Agent Communication Language (ACL) is a collection of speech-act-like message types, with agreed-upon semantics, which facilitate the knowledge and information exchange between software agents. | [
996,
1552,
2629
] | Train |
790 | 4 | Ten Myths of Multimodal Interaction | [
1778,
2069,
2788,
2955
] | Train |
791 | 3 | Maintaining Temporal Views Over Non-Temporal Information Sources For Data Warehousing An important use of data warehousing is to provide temporal views over the history of source data that may itself be non-temporal. While recent work in view maintenance is applicable to data warehousing, only non-temporal views have been considered. In this paper, we introduce a framework for maintaining temporal views over non-temporal information sources in a data warehousing environment. We describe an architecture for the temporal data warehouse that automatically maintains temporal views over non-temporal source relations, and allows users to ask temporal queries using these views. Because of the dimension of time, a materialized temporal view may need to be updated not only when source relations change, but also as time advances. We present incremental techniques to maintain temporal views for both cases, and outline the implementation of our approach in the WHIPS warehousing prototype at Stanford. 1 Introduction A data warehouse is a repository for efficient querying ... | [
1894,
3123
] | Train |
792 | 0 | Systems Management In Concurrent Engineering Using Intelligent Software Agents Intelligent software agents are used in frameworks where large number of experts need to interact in a project concurrently as in the projects taken up by the aerospace industry. We describe one such framework and discuss an intelligent software agent to manage the systems design in such an environment. What makes the problem interesting is the existence of other intelligent agents in the framework that are responsible for various other tasks as well as other human users. Our Systems Design Management Agent (SDMA) uses its own domain knowledge to interact with the other agents and recommend strategies and policies. We take one example task of systems design management, risk management, and discuss how it is performed by the SDMA in detail. 1 Introduction Using intelligent software agents within the concurrent engineering paradigm has received attention from several research groups. The idea is to respond to the increased information and coordination demands of concurrent engineering p... | [
1107
] | Train |
793 | 1 | Active Markov Localization for Mobile Robots Localization is the problem of determining the position of a mobile robot from sensor data. Most existing localization approaches are passive, i.e., they do not exploit the opportunity to control the robot's effectors during localization. This paper proposes an active localization approach. The approach is based on Markov localization and provides rational criteria for (1) setting the robot's motion direction (exploration), and (2) determining the pointing direction of the sensors so as to most efficiently localize the robot. Furthermore, it is able to deal with noisy sensors and approximative world models. The appropriateness of our approach is demonstrated empirically using a mobile robot in a structured office environment. Key words: Robot Position Estimation, Autonomous Service Robots 1 Introduction To navigate reliably in indoor environments, a mobile robot must know where it is. Over the last few years, there has been a tremendous scientific interest in algorithms for estimating ... | [
213,
548,
641,
1093,
1565,
2190,
2999
] | Test |
794 | 5 | Three New Algorithms for Projective Bundle Adjustment with Minimum Parameters Bundle adjustment is a technique used to compute the maximum likelihood estimate of structure and motion from image feature correspondences. It practice, large non-linear systems have to be solved, most of the time using an iterative optimization process starting from a sub-optimal solution obtained by using linear methods. The behaviour, in terms of convergence, and the computational cost of this process depend on the parameterization used to represent the problem, i.e. of structure and motion. | [
1417
] | Test |
795 | 0 | Non-Supervised Sensory-Motor Agents Learning This text discusses a proposal for creation and destruction of neurons based on the sensory-motor activity. This model, called sensory-motor schema, is used to define a sensory-motor agent as a collection of activity schemata. The activity schema permits a useful distribution of neurons in a conceptual space, creating concepts based on action and sensation. Such approach is inspired in the theory of the Swiss psychologist and epistemologist Jean Piaget, and intends to make explicit the account of the processes of continuous interaction between sensory-motor agents and their environments when agents are producing cognitive structures. 1. Introduction The notion of an autonomous agent plays a central role in contemporaneous research on Artificial Intelligence [3]. Cognitive agents are based on symbolic processing mechanisms. Reactive agents are based on alternative computational mechanisms like neural networks, analogic processing, etc. The alternative approach using autonomous agents b... | [
2858
] | Test |
796 | 4 | Using Context as a Crystal Ball: Rewards and Pitfalls This paper discusses some of the potential rewards and pitfalls that can await designers wishing to incorporate context-awareness [Schilit,94][Brown,97] into interactive systems. Many of the issues are described in anecdotal form, based on our experiences developing and evaluating the context-aware GUIDE system [Cheverst,99][Cheverst,00]. | [
293,
2855
] | Train |
797 | 3 | Simultaneous Proxy Evaluation The Simultaneous Proxy Evaluation (SPE) architecture is designed to evaluate multiple web proxies in parallel using object requests which are duplicated and passed to each proxy. The SPE architecture reduces problems of unrealistic test environments, dated and/or inappropriate workloads, and is additionally applicable to contentbased prefetching proxies. It is intended to measure byte and object hit rates, client-perceived latencies, and cache consistency. We characterize a space of proxy evaluation methodologies and place this architecture within it. 1 Introduction This paper presents a new architecture for the evaluation of proxy caches. Initially, it grew out of research in techniques for prefetching in web caches. In particular, we found that existing mechanisms for the evaluation of proxy caches were not well suited to prefetching systems. Objective evaluation is paramount to all research, whether applied or academic. Since this is certainly relevant when exploring various approac... | [
2518
] | Train |
798 | 4 | POWERVIEW - Using information links and information views to navigate and visualize information on small displays . PowerView is a PDA application designed to support people with situational information, primarily during conversations and meetings with other people. PowerView was designed to address a number of issues in interface design concerning both information visualization and interaction on small, mobile devices. In terms of information visualization, the system was required to provide the user with a single integrated information system that enabled quick access to related information once an object of interest had been selected. In terms of interaction, the system was required to enable easy and efficient information retrieval, including single-handed use of the device. These problems were addressed by introducing Information Links and Information Views. An evaluation of the application against the standard application suite bundle of the PDA, a Casio Cassiopeia E-11, proved the interfaces equivalent in usability even though the PowerView application uses a novel interface par... | [
1500,
2064
] | Train |
799 | 1 | Learning Markov Processes this article, we restrict our attention to discrete time dynamical systems.) Typically we do not know the exact dynamics of the system, so instead we consider a probabilistic state transition function: P (X t+1 jX t ). Such a probabilistic formulation will be particularly useful when we try to learn the model from data. The state space, might be discrete (nite) or continuous (innite). For example, we might just try to predict the probability that a stock goes up or down, in which case = f"; #g; more ambitiously, we might try to predict its expected value, in which case = IR. In general, the state is a vector of state variables, which we can partition into three kinds: input variables (ones which we can control), output variables (ones which we can observe), and hidden or latent variables (internal variables which we cannot directly control or observe); we shall denote these by U t , Y t and X t respectively. See Figure 1. In this article, we shall consider how to learn models of this kind. We start by considering the special case in To be published in The Encyclopedia of Cognitive Science, Macmillan, 2002 1 X1 X2 X3 Y1 Y2 Y3 U1 U2 . . . Figure 1: A generic discrete-time dynamical system represented as a dynamic Bayesian network (DBN) (see BAYESIAN BELIEF NETWORKS for a denition). U t is the input, X t is the hidden state, and Y t is the output. Shaded nodes are observed, clear nodes are hidden. Square nodes are xed inputs (controls), round nodes are random variables. Notice how what we see, Y t , may depend on the actions that we take, U t : this can be used to model active perception. X1 X2 X3 X4 . . . X1 X2 X2 X3 X3 X4 Figure 2: Converting a second order Markov model (top) into a rst order Markov model (bottom). which all variables are observed (i.e.... | [
2463
] | Test |
800 | 3 | Exploiting Versions for Handling Updates in Broadcast Disks Recently, broadcasting has attracted considerable attention as a means of disseminating information to large client populations in both wired and wireless settings. In this paper, we exploit versions to increase the concurrency of client transactions in the presence of updates. We consider three alternative mediums for storing versions: (a) the air: older versions are broadcast along with current data, (b) the client's local cache: older versions are maintained in cache, and (c) a local database or warehouse at the client: part of the server's database is maintained at the client in the form of a multiversion materialized view. The proposed techniques are scalable in that they provide consistency without any direct communication from clients to the server. Performance results show that the overhead of maintaining versions can be kept low, while providing a considerable increase in concurrency. 1 Introduction While traditionally data are delivered from servers to clients on demand, a... | [
381,
768,
954,
1047,
2543
] | Train |
801 | 2 | Knowledge-based Wrapper Generation by Using XML Information extraction is the process of recognizing the particular fragments of a document that constitute its core semantic content. However, most previous information extraction systems were not effective for real-world information sources due to difficulties in acquiring and representing useful domain knowledge and in dealing with structural heterogeneity inherent in different sources. In order to resolve these problems, this paper proposes a scheme of knowledge-based wrapper generation for semi-structured and labeled documents. The implementation of an agent-oriented information extraction system, XTROS, is described. In contrast with previous wrapper learning agents, XTROS represents both the domain knowledge and the wrappers by XML documents to increase modularity, flexibility, and interoperability among multiple parties. XTROS also facilitates simpler implementation of the wrapper generator by exploiting XML parsers and interpreters. XTROS shows good performance on several Web sites in the domain of real estates, and it is expected to be easily adaptable to different domains by plugging in appropriate XML-based domain knowledge. 1 | [
1090,
2068,
3053
] | Train |
802 | 3 | The Persistent Cache: Improving OID Indexing in Temporal Object-Oriented Database Systems In a temporal OODB, an OID index (OIDX) is needed to map from OID to the physical location of the object. In a transaction time temporal OODB, the OIDX should also index the object versions. In this case, the index entries, which we call object descriptors (OD), also include the commit timestamp of the transaction that created the object version. The OIDX in a non-temporal OODB only needs to be updated when an object is created, but in a temporal OODB, the OIDX have to be updated every time an object is updated. We have in a previous study shown that this can be a potential bottleneck, and in this report, we present the Persistent Cache (PCache), a novel approach which reduces the index update and lookup costs in temporal OODBs. In this report, we develop a cost model for the PCache, and use this to show that the use of a PCache can reduce the average access cost to only a fraction of the cost when not using the PCache. Even though the primary context of this report is OID indexing in ... | [
1441
] | Train |
803 | 3 | Description Logics for Information Integration Information integration is the problem of combining the data residing at different, heterogeneous sources, and providing the user with a unified view of these data, called mediated schema. The mediated schema is therefore a reconciled view of the information, which can be queried by the user. It is the task of the system to free the user from the knowledge on where data are, and how data are structured at the sources. | [
2990
] | Train |
804 | 4 | Reaching for Objects in VR Displays: Lag and Frame Rate This article reports the results from three experimental studies of reaching behavior in a head-coupled stereo display system with a hand-tracking subsystem for object selection. It is found that lag in the head-tracking system is relatively unimportant in predicting performance, whereas lag in the hand-tracking system is critical. The effect of hand lag can be modeled by means of a variation on Fitts ’ Law with the measured system lag introduced as a multiplicative variable to the Fitts ’ Law index of difilculty. This means that relatively small lags can cause considerable degradation in performance if the targets are small. Another finding is that errors are higher for movement in and out of the screen, as compared to movements in the plane of the screen, and there is a small (10’%) time penalty for movement in the Z direction in all three experiments. Low frame rates cause a degradation in performance; however, this can be attributed to the lag which is caused by low frame rates, particularly if double buffering is used combined with early sampling of the hand-tracking device. | [
1435,
2185
] | Test |
805 | 5 | Scalability In Formal Concept Analysis This paper presents the results of experiments carried out with a set of 4,000 medical discharge summaries in which were recognised 1,962 attributes from the Unified Medical Language System (UMLS). In this domain, the objects are medical documents (4,000) and the attributes are UMLS terms extracted from the documents (1,962). When Formal Concept Analysis is used to iteratively analyse and visualize this data, complexity and scalability become critically important. Although the amount of data used in this experiment is small compared with the size of primary memory in modern computers, the results are still important since the probability distributions which determine the efficiencies are likely to remain stable as the size of the data is increased. Our work presents two outcomes, firstly we present a methodology for exploring knowledge in text documents using Formal Concept Analysis by employing conceptual scales created as the result of direct manipulation of a line diagram. The conceptual scales lead to small derived purified contexts that are represented using nested line diagrams. Secondly, we present an algorithm for the fast determination of purified contexts from a compressed representation of the large formal context. Our work draws on existing encoding and compression techniques to show how rudimentary data analysis can lead to substantial efficiency improvements to knowledge visualisation. c fl 1993 Blackwell Publishers, 238 Main Street, Cambridge, MA 02142, USA, and 108 Cowley Road, Oxford, OX4 1JF, UK. SCALABILITY IN FORMAL CONCEPT ANALYSIS 3 | [
2787
] | Test |
806 | 3 | Synchronizing a database to Improve Freshness In this paper we study how to refresh a local copy of an autonomous data source to maintain the copy up-to-date. As the size of the data grows, it becomes more difficult to maintain the copy "fresh," making it crucial to synchronize the copy effectively. We define two freshness metrics, change models of the underlying data, and synchronization policies. We analytically study how effective the various policies are. We also experimentally verify our analysis, based on data collected from 270 web sites for more than 4 months, and we show that our new policy improves the "freshness" very significantly compared to current policies in use. | [
40,
1177,
1750,
2503,
2615,
2739
] | Train |
807 | 0 | Programmable Agents for Flexible QoS Management in IP Networks Network programmability seems to be a promising solution to network management and quality of service (QoS) control. Software mobile-agents technology is boosting the evolution toward application-level control of network functionalities. Code may be deployed in the network dynamically and on demand for the benefit of applications or application classes. Agents support a dynamic distribution of control and management functions across networks, thus increasing flexibility and efficiency. We propose to use mobile-agent technology to overcome some of the problems inherent in current Internet technology. We focus our attention to QoS monitoring, being locally significant in network subdomains, and realize a QoS management strategy in response to variations of user, customer of application requirements, and of the network state. We describe our experience and the results obtained from our testbed, where software agents are instantiated, executed, migrated, and suspended in order to implement flexible QoS management in IP networks. | [
685
] | Train |
808 | 0 | A Knowledge-Based Approach for Designing Intelligent Team Training Systems This paper presents a knowledge approach to designing team training systems using intelligent agents. We envision a computer-based training system in which teams are trained by putting them through scenarios, which allow them to practice their team skills. There are two important roles that intelligent agents can play; these are virtual team members, and tutors. To carry out these functions, these agents must be equipped with an understanding of the task domain, the team structure, the selected decision-making process and their beliefs about other team members' mental states. Even though existing agent teamwork models incorporate many of the elements listed above, they have not focused on analyzing information needs of team members to support proactive agent interactions. To encode the team knowledge, we have developed a representation language, based on the BDI model, called MALLET. A Petri Net model of an individual agent's plans and information needs can be derived from the role des... | [
1614,
1840,
1923,
2364
] | Train |
809 | 1 | A Framework for Programming Embedded Systems: Initial Design and Results This paper describes CES, a proto-type of a new programming language for robots and other embedded systems, equipped with sensors and actuators. CES contains two new ideas, currently not found in other programming languages: support of computing with uncertain information, and support of adaptation and teaching as a means of programming. These innovations facilitate the rapid development of software for embedded systems, as demonstrated by two mobile robot applications. This research is sponsoredin part by DARPA via AFMSC (contract number F04701-97-C-0022), TACOM (contract number DAAE07-98-C-L032), and Rome Labs (contract number F30602-98-2-0137). The views and conclusions contained in this document are those of the author and should not be interpreted as necessarily representing official policies or endorsements, either expressed or implied, of DARPA, AFMSC, TACOM, Rome Labs, or the United States Government. Keywords: Artificial intelligence, embedded system, machine learning, mobil... | [
2100,
2999
] | Test |
810 | 4 | Improving Interaction with Virtual Environments Introduction Virtual environments (VEs) provide a computer-based interface to a real-life or abstract space, using 3D graphics and 3D interaction techniques. VEs represent a novel interface style which offers new possibilities and challenges to human-computer interface design. However, studies of the design of VEs (Kaur et al., 1996) show that designers lack a coherent approach to design, especially interaction design. Designers appear to be pre-occupied with difficult technical issues and think little about supporting user interaction. However, major interaction problems have been found with current VEs, such as disorientation, perceptual misjudgements and difficulty finding and understanding available interactions (McGovern, 1993; COVEN, 1997). These common problems have been known to result in user frustration and a low usability and acceptability for the VE (Kaur et al., 1996; Miller 1994). Guidance is needed on interaction design for VEs to avoid such usability problems. | [
1585,
2649,
2717
] | Train |
811 | 0 | Formal Specification and Design of Mobile Systems Termination detection, a classical problem in distributed computing, is revisited in the new setting provided by the emerging mobile computing technology. A simple solution tailored for use in ad hoc networks is employed as a vehicle for demonstrating the applicability of formal requirements and design strategies to the new field of mobile computing. The approach is based on well understood techniques in specification refinement, but the methodology is tailored to mobile applications and helps designers address novel concerns such as the mobility of hosts, transient interactions, and specific coordination constructs. The proof logic and programming notation of Mobile UNITY provide the intellectual tools required to carry out this task. | [
580
] | Train |
812 | 0 | Multi-Agent Reinforcement Learning: Weighting and Partitioning This paper addresses weighting and partitioning in complex reinforcement learning tasks, with the aim of facilitating learning. The paper presents some ideas regarding weighting of multiple agents and extends them into partitioning an input/state space into multiple regions with differential weighting in these regions, to exploit differential characteristics of regions and differential characteristics of agents to reduce the learning complexity of agents (and their function approximators) and thus to facilitate the learning overall. It analyzes, in reinforcement learning tasks, different ways of partitioning a task and using agents selectively based on partitioning. Based on the analysis, some heuristic methods are described and experimentally tested. We find that some off-line heuristic methods performed the best, significantly better than single-agent models. Keywords: weighting, averaging, neural networks, partitioning, gating, reinforcement learning, 1 Introduction Multiple ag... | [
1190
] | Train |
813 | 5 | From Markov Random Fields to Associative Memories and Back: Spin-Glass Markov Random Fields this paper we propose a fully connected energy function for Markov Random Field (MRF) modeling which is inspired by Spin-Glass Theory (SGT). Two major tasks in MRF modeling are how to define the neighborhood system for irregular sites and how to choose the energy function for a proper encoding of constraints. The proposed energy function offers two major advantages that makes it possible to avoid MRF modeling problems in the case of irregular sites. First, full connectivity makes the neighborhood definition irrelevant, and second, the energy function is defined independently of the considered application. A basic assumption in SGT is the infinite dimension of the configuration space in which the energy is defined; the choice of a particular energy function, which depends on the scalar product between configurations, allows us to use a kernel function in the energy formulation; this solves the problem of high dimensionality and makes it possible to use SGT results in an MRF framework. We call this new model Spin Glass- - Markov Random Field (SG-MRF). Experiments on textures and objects database show the correctness and effectiveness of the proposed model | [] | Train |
814 | 5 | HICAP: An Interactive Case-Based Planning Architecture and its Application to Noncombatant Evacuation Operations This paper describes HICAP (Hierarchical Interactive Case-based Architecture for Planning), a general purpose planning architecture that we have developed and applied to assist military commanders and their staff with planning NEOs (Noncombatant Evacuation Operations). HICAP integrates a hierarchical task editor, HTE, with a conversational case-based planning tool, NaCoDAE/HTN. In this application, HTE maintains an agenda of tactical planning tasks that, according to the guidelines indicated by military doctrine, must be addressed in a NEO plan. It also supports several bookkeeping tasks, which are crucial for large-scale planning tasks that differ greatly among different NEO operations. Military planning personnel select a task to decompose from HTE and then use NaCoDAE/HTN to interactively refine it into an operational plan by selecting and applying cases, which represent task decompositions from previous NEO operations. Thus, HICAP helps commanders by using previous experience to fo... | [
2258
] | Train |
815 | 2 | Learning to Construct Knowledge Bases from the World Wide Web The World Wide Web is a vast source of information accessible to computers, but understandable only to humans. The goal of the research described here is to automatically create a computer understandable knowledge base whose content mirrors that of the World Wide Web. Such a knowledge base would enable much more effective retrieval of Web information, and promote new uses of the Web to support knowledge-based inference and problem solving. Our approach is to develop a trainable information extraction system that takes two inputs. The first is an ontology that defines the classes (e.g., company, person, employee, product) andrelations (e.g., employed by, produced by) of interest when creating the knowledge base. The second is a set of training data consisting of labeled regions of hypertext that represent instances of these classes and relations. Given these inputs, the system learns to extract information from other pages and hyperlinks on the Web. This article describes our general approach, several machine learning algorithms for this task, and promising initial results with a prototype system that has created a knowledge base describing university people, courses, and research projects. | [
301,
702,
1386,
1478,
2068,
2133,
2342,
2391,
2765,
2961
] | Train |
816 | 1 | Graded Learning for Object Detection Our goal is to detect all instances of a generic object class, such as a face, in greyscale scenes. The design of the algorithm is motivated by computational efficiency. The search is coarse-to-fine in both the exploration of poses and the representation of the object class. Starting from training examples, we recursively learn a hierarchy of spatial arrangements of edge fragments, graded by their size (sparsity). The arrangements have no a priori semantic or geometric interpretation. Instead, they are selected to be "decomposable": Each can be split into two correlated subarrangements, each of which can be further divided, etc. As a result, the probability of an arrangement of size k appearing on an object instance decays slowly with k. We demonstrate this both theoretically and in experiments in which detection means finding a sufficient number of arrangements of various sizes. 1 Introduction Starting with a training set of examples of a generic object class (e.g., "face"), our goal... | [
1263,
2694
] | Train |
817 | 3 | Dynamic Update Cube for Range-Sum Queries A range-sum query is very popular and becomes important in finding trends and in discovering relationships between attributes in diverse database applications. It sums over the selected cells of an OLAP data cube where target cells are decided by the specified query ranges. The direct method to access the data cube itself forces too many cells to be accessed, therefore it incurs a severe overhead. The response time is very crucial for OLAP applications which need interactions with users. In the recent dynamic enterprise environment, data elements in the cube are frequently changed. The response time is affected in such an environment by the update cost as well as the search cost of the cube. In this paper, we propose an efficient algorithm to reduce the update cost significantly while maintaining reasonable search efficiency, by using an index structure called the -tree. In addition, we propose a hybrid method to provide either an approximate result or a precise one to reduce the overall cost of queries. It is useful for various applications that need a quick approximate answer rather than an accurate one, such as decision support systems. 1. | [
2095
] | Test |
818 | 1 | Fuzzy Argumentation and Extended Logic Programming We define a fuzzy version of extended logic programming under well-founded semantics with explicit negation. We develop a fuzzy fixpoint argumentation semantics and an efficient, top-down, goal-directed proof procedure. We show that the procedure is sound and complete with respect to the argumentation semantics. | [
1260
] | Train |
819 | 0 | Concepts and Architecture of a Security-Centric Mobile Agent Server Mobile software agents are software components that are able to move in a network. They are often considered as an attractive technology in electronic commerce applications. Although security concerns prevail. In this paper we describe the architecture and concepts of the SeMoA server -- a runtime environment for Java-based mobile agents. Its architecture has a focus on security and easy extendability, and offers a framework for transparent content inspection of agents by means of filters. We implemented filters that handle agent signing and authentication as well as selective encryption of agent contents. Filters are applied transparently such that agents need not be aware of the security services provided by the server. | [
2715,
2846,
3114
] | Validation |
820 | 0 | Algorithms for Optimizing Leveled Commitment Contracts In automated negotiation systems consisting of self-interested agents, contracts have traditionally been binding. Leveled commitment contracts—i.e. contracts where each party can decommit by paying a predetermined penalty were recently shown to improve Pareto efficiency even if agents rationally decommit in Nash equilibrium using inflated thresholds on how good their outside offers must be before they decommit. This paper operationalizes the four leveled commitment contracting protocols by presenting algorithms for using them. Algorithms are presented for computing the Nash equilibrium decommitting thresholds and decommitting probabilities given the contract price and the penalties. Existence and uniqueness of the equilibrium are analyzed. Algorithms are also presented for optimizing the contract itself (price and penalties). Existence and uniqueness of the optimum are analyzed. Using the algorithms we offer a contract optimization service on the web as part of ('Mediator, our next generation electronic commerce server. Finally, the algorithms are generalized to contracts involving more than two agents. 1 | [
1337,
1346,
2546
] | Validation |
821 | 3 | Query Processing in Relationlog . Relationlog is a persistent deductive database system that supports eective storage, ecient access and inference of large amounts of data with complex structures. In this paper, we describe query processing in the Relationlog system. In particular, we illustrate the extended semi-naive and magic-set techniques used in Relationlog. 1 Introduction During the past decades, the nested relational and complex object models [1, 5, 11, 12, 15, 18, 21, 22] were developed to extend the applicability of the traditional relational model to more complex, non-business applications such as CAD, image processing and text retrieval [2]. Another important direction of intense research has been in using a logic programming based language Datalog [8, 23] as a database query language. Such a language provides a simple and natural way to express queries on a relational database and is more expressive than the traditional relational languages. In the past several years, there have been some eorts... | [
418,
2083
] | Validation |
822 | 3 | ADOME: An Advanced Object Modelling Environment ADOME, ADvanced Object Modeling Environment, an approach to integrating data and knowledge management based on object-oriented technology, is presented. Next generation information systems will require more flexible data modelling capabilities than those provided by current object-oriented DBMSs. In particular, integration of data and knowledge management capabilities will become increasingly important. In this context, ADOME provides versatile role facilities that serve as "dynamic binders" between data objects and production rules, thereby facilitating flexible data and knowledge management integration. A prototype that implements this mechanism and the associated operators has been constructed on top of a commercial object-oriented DBMS and a rule base system. Index Terms: Object modeling, knowledge semantics, dynamic roles, object-oriented databases, nextgeneration information systems 1 Introduction Increasingly, organizations require more intelligent information management. In o... | [
268,
3064
] | Train |
823 | 3 | A Web Agent for the Maintenance of a Database of Academic Contacts this paper to present all details of the confidence rules for Academia, however these are fully discussed in (Magnanelli 1997). It is important to point out that we consider the information in the Web not only as free to use but also as true and updated. The agent is unable to detect that information is wrong in the case that the correct information is not available. 6 User Interaction | [
436
] | Train |
824 | 0 | On splitting and Cloning Agents Embedded with cloning mechanisms, an agent can balance its own loads by discharging computing tasks to its clones when it is over-loaded. In addition, it's more reasonable to transfer the smarter, smaller clones of an agent rather than the bulky agent itself in mobile computing. In this paper, a simple BDI agent model is formally established. Using this model, the semantics of constructing new agents by inheritance and self-identifying behavior of existing agents are precisely de ned. Four kinds of cloning mechanisms are identi ed, the properties of each cloning mechanism and the relationships in between are studied, and some implementation issues are also discussed. | [
3168
] | Train |
825 | 2 | Information Retrieval on the World Wide Web and Active Logic: A Survey and Problem Definition As more information becomes available on the World Wide Web (there are currently over 4 billion pages covering most areas of human endeavor), it becomes more difficult to provide effective search tools for information access. Today, people access web information through two main kinds of search interfaces: Browsers (clicking and following hyperlinks) and Query Engines (queries in the form of a set of keywords showing the topic of interest). The first process is tentative and time consuming and the second may not satisfy the user because of many inaccurate and irrelevant results. Better support is needed for expressing one's information need and returning high quality search results by web search tools. There appears to be a need for systems that do reasoning under uncertainty and are flexible enough to recover from the contradictions, inconsistencies, and irregularities that such reasoning involves. | [
488,
608,
1016,
1021,
1838,
2283,
2532,
2610,
2650,
2971
] | Validation |
826 | 3 | DyDa: Dynamic Data Warehouse Maintenance in a Fully Concurrent Environment Data warehouse is an emerging technology to support high-level decision making by gathering data from several distributed information sources into one repository. In dynamic environments, data warehouses must be maintained in order to stay consistent with the underlying sources. Recently proposed view maintenance algorithms tackle the problem of data warehouse maintenance under concurrent source data updates.While the view synchronization is to handle non-concurrent source schema changes. However, the concurrency between interleaved schema changes and data updates still remain unexplored problems. In this paper, we propose a solution framework called DyDa that successfully addresses this problem. The DyDa framework detects concurrent schema changes by a broken query scheme and conicting concurrent data updates by a local timestamp scheme. A fundamental idea of the DyDa framework is the development of a two-layered architecture that separates the concerns for concurrent data updates and concurrent schema changes handling without imposing any restrictions on the sourse update transactions. At the lower level of the framework, it employs a local compensation algorithm to handle concurrent data updates, and a metadata name mapping strategy to handle concurrent source rename operations. At the higher level, it addresses the problem of concurrent source drop operations. For the latter problem, we design a strategy for the detection and correction of such concurrency and nd an executable plan for the aected updates. We then develop a new view adaption algorithm, called Batch-VA for execution of such plan to incrementally adapt the view. Put together, these algorithms are the rst to provide a complete solution to data warehouse management in a fully concurrent environment.... | [
1018,
1463
] | Train |
827 | 0 | Antisocial Agents and Vickrey Auctions In recent years auctions have become more and more important in the field of multiagent systems as useful mechanisms for resource allocation and task assignment. In many cases the Vickrey (second-price sealed-bid) auction is used as a protocol that prescribes how the individual agents have to interact in order to come to an agreement. We show that the Vickrey auction, despite its theoretical benefits, is inappropriate if "antisocial" agents participate in the auction process. More specifically, an antisocial attitude for economic agents that makes reducing the profit of competitors their main goal besides maximizing their own profit is introduced. Under this novel condition, agents need to deviate from the dominant truth-telling strategy. This paper presents a strategy for bidders in repeated Vickrey auctions who are intending to inflict losses to fellow agents in order to be more successful, not in absolute measures, but relatively to the group of bidders. The strategy is evaluated in a simple task allocation scenario. | [
1418,
1561,
2602
] | Train |
828 | 4 | Practical Guidelines for the Readability of IT-architecture Diagrams This paper presents the work done to establish guidelines for the creation of readable IT-architecture diagrams and gives some examples of guidelines and some examples of improved diagrams. These guidelines are meant to assist practicing IT-architects in preparing the diagrams to communicate their architectures to the various stakeholders. Diagramming has always been important in information technology (IT), but the recent interest in ITarchitecture, the widespread use of software and developments in electronic communication, make it necessary to again look at the rt of making diagrams'for this particular class and its users. The guidelines indicate how various visual attributes, like hierarchy, layout, color, form, graphics, etc. can contribute to the readability of IT-architecture diagrams. The emphasis is on the outward appearance of diagrams. Some additional support is given for the thinking/reasoning processes while designing or using a set of diagrams and an attempt is made to arrive at a rationale of these guidelines. An evaluation process has been performed with three groups of practicing IT-architects. The outcome of this evaluation is presented. This work is part of a more comprehensive research project on "Visualisation of IT- architecture". | [
939
] | Validation |
829 | 0 | Manufacturing Systems Integration and Agility: Can Mobile Agents Help? Mobile code is being championed as a solution to a plethora of software problems. This paper investigates whether Mobile Agents and Mobile Objects support improved system integration and agility in the manufacturing domain. We describe two systems built to support the Sales Order Process of a distributed manufacturing enterprise, using IBM's Aglets Software Development Kit. The Sales Order Process model and the requirements for agility used as the basis for these implementations are derived from data collected in an industrial case study. Both systems are evaluated using the Goal/Question/Metric methodology. Two new metrics for Semantic Alignment and Change Capability are presented and used to evaluate each system with respect to the degree of system agility supported. The work described provides evidence that both Mobile Agent and Mobile Object systems have inherent properties that can be used to build agile distributed systems. Further, Mobile Agents with their additional autonomy... | [
1035
] | Validation |
830 | 2 | Information Extraction by Text Classification: Corpus Mining for Features This paper describes a method for building an Information Extraction (IE) system using standard text classification machine learning techniques, and datamining for complex features on a large corpus of example texts that are only superficially annotated. We have successfully used this method to build an IE system (Textractor) for job advertisements. 1. Introduction For rapid development of an Information Extraction system in a large new domain, the usual methods of semicorpusbased hand-crafting of extraction rules are often simply too laborious. Therefore one must turn to the use of machine learning techniques and try to induce the knowledge needed for extraction from annotated training samples. Techniques for the induction of extraction rules are e.g. described by (Freitag, 1998 | [
846
] | Train |
831 | 1 | Rough Set Theory: A Data Mining Tool for Semiconductor Manufacturing The growing volume of information poses interesting challenges and calls for tools that discover properties of data. Data mining has emerged as a discipline that contributes tools for data analysis, discovery of new knowledge, and autonomous decisionmaking. In this paper, the basic concepts of rough set theory and other aspects of data mining are introduced. The rough set theory offers a viable approach for extraction of decision rules from data sets. The extracted rules can be used for making predictions in the semiconductor industry and other applications. This contrasts other approaches such as regression analysis and neural networks where a single model is built. One of the goals of data mining is to extract meaningful knowledge. The power, generality, accuracy, and longevity of decision rules can be increased by the application of concepts from systems engineering and evolutionary computation introduced in this paper. A new rule-structuring algorithm is proposed. The concepts presented in the paper are illustrated with examples. | [
1,
998
] | Validation |
832 | 1 | Support Vector Machines for Classification and Regression The problem of empirical data modelling is germane to many engineering applications. In empirical data modelling a process of induction is used to build up a model of the system, from which it is hoped to deduce responses of the system that have yet to be observed. Ultimately the quantity and quality of the observations govern the performance of this empirical model. By its observational nature data obtained is finite and sampled; typically this sampling is non-uniform and due to the high dimensional nature of the problem the data will form only a sparse distribution in the input space. Consequently the problem is nearly always ill posed (Poggio et al., 1985) in the sense of Hadamard (Hadamard, 1923). Traditional neural network approaches have suffered difficulties with generalisation, producing models that can overfit the data. This is a consequence of the optimisation algorithms used for parameter selection and the statistical measures used to select the ’best’ model. The foundations of Support Vector Machines (SVM) have been developed by Vapnik (1995) and are gaining popularity due to many attractive features, and promising empirical performance. The formulation embodies the Structural Risk Minimisation (SRM) principle, which has been shown to be superior, (Gunn et al., 1997), to traditional Empirical Risk Minimisation (ERM) principle, employed by conventional neural networks. SRM minimises an upper bound on the expected risk, as opposed to ERM that minimises the error on the training data. It is this difference which equips SVM with a greater ability to generalise, which is the goal in statistical learning. SVMs were developed to solve the classification problem, but recently they have been extended to the domain of regression problems (Vapnik et al., 1997). In the literature the terminology for SVMs can be slightly confusing. The term SVM is typically used to describe classification with support vector methods and support vector regression is used to describe regression with support vector methods. In this report the term SVM will refer to both classification and regression methods, and the terms Support Vector Classification (SVC) and Support Vector Regression (SVR) will be used for specification. This section continues with a brief introduction to the structural risk | [
2410,
2675
] | Train |
833 | 1 | Learning with Deictic Representation Most reinforcement learning methods operate on propositional representations of the world state. Such representations are often intractably large and generalize poorly. Using a deictic representation is believed to be a viable alternative: they promise generalization while allowing the use of existing reinforcement-learning methods. Yet, there are few experiments on learning with deictic representations reported in the literature. In this paper we explore the effectiveness of two forms of deictic representation and a naïve propositional representation in a simple blocks-world domain. We find, empirically, that the deictic representations actually worsen performance. We conclude with a discussion of possible causes of these results and strategies for more effective learning in domains with objects. 1 | [
1467,
1571
] | Validation |
834 | 2 | Representation of Electronic Mail Filtering Profiles: A User Study Electronic mail offers the promise of rapid communication of essential information. However, electronic mail is also used to send unwanted messages. A variety of approaches can learn a profile of a user's interests for filtering mail. Here, we report on a usability study that investigates what types of profiles people would be willing to use to filter mail. Keywords Mail Filtering; User Studies 1. INTRODUCTION While electronic mail offers the promise of rapid communication of essential information, it also facilitates transmission of unwanted messages such as advertisements, solicitations, light bulb jokes, chain letters, urban legends, etc. Software that automatically sorts mail into categories (e.g., junk, talk announcements, homework questions) would help automate the process of sorting through mail to prioritize messages or suggest actions (such as deleting junk mail or forwarding urgent messages to a handheld device). Such software maintains a profile of the user's interests. ... | [
2424
] | Train |
835 | 2 | Context Filters for Document-Based Information Filtering In this paper we propose a keyPhrase-sense disambiguation methodology called "context filters" for use in keyPhrase based information filtering systems. A context filter finds whether an input keyPhrase has occurred in the required context. Context filters consider various factors of ambiguity. Some of these factors are special to information filtering and they are handled in a structured fashion. The proposed context filters are very comprehensibile. Context filters consider varieties of contexts which are not considered in existing word-sense disambiguation methods but these are all needed for information filtering. The ideas on context filters that we report in this paper form important elements of an Instructible Information Filtering Agent that we are developing. 1. Introduction Information filtering is the process of separating out irrelevant documents from relevant ones. Its importance has motivated several researchers to develop software agents such as SIFT, InfoScan, iAgent, ... | [
2455,
2477
] | Validation |
836 | 1 | A Machine Learning Approach to Building Domain-Specific Search Engines Domain-specific search engines are becoming increasingly popular because they offer increased accuracy and extra features not possible with general, Web-wide search engines. Unfortunately, they are also difficult and timeconsuming to maintain. This paper proposes the use of machine learning techniques to greatly automate the creation and maintenance of domain-specific search engines. We describe new research in reinforcement learning, text classification and information extraction that enables efficient spidering, populates topic hierarchies, and identifies informative text segments. Using these techniques, we have built a demonstration system: a search engine for computer science research papers available at www.cora.justresearch.com. 1 Introduction As the amount of information on the World Wide Web grows, it becomes increasingly difficult to find just what wewant. While general-purpose search engines suchas AltaVista and HotBot offer high coverage, they often provi... | [
255,
603,
759,
1016,
1386,
1532,
1543,
1843,
2371
] | Train |
837 | 4 | Gaia: Enabling Active Spaces Ubiquitous computing promotes physical spaces with hundreds of specialized embedded devices that increase our productivity, alleviate some specific everyday tasks and provide new ways of interacting with the computational environment. Personal computers lose the focus of attention due to the fact that the computational environment is spread across the physical space. Therefore, the users' view of the computational environment is finally extended beyond the physical limits of the computer. Physical spaces become computer systems, or in other terms, Active Spaces. However, these Active Spaces require novel system software capable of seamlessly coordinating their hidden complexity. Our goal is to extend the model provided by current computer systems to allow interaction with physical spaces and their contained entities (physical and virtual) by means of a single abstraction called Active Space. 1. Introduction Ubiquitous computing promotes the proliferation of embedded devices specializ... | [
422,
1649,
2282
] | Train |
838 | 4 | Providing an Embedded Software Environment for Wireless PDAs . The use of wireless Pdas is foreseen to outrun the one of Pcs in the near future. However, for this to actually happen, adequate software environments must be devised in order to allow the execution of various types of applications. This paper introduces the base features of such an environment, which is a customizable Jvm-based middleware. In particular, the middleware platform embeds services for appropriate resource management and for supporting novel Pda-oriented applications. 1 Introduction The use of wireless Personal Digital Assistant (Pda) devices is foreseen to outrun the one of Pcs in the near future. However, for this to actually happen, there is still the need to devise adequate software and hardware platforms. The use of Pdas should be as convenient as the one of Pcs and in particular must not overly restrict the applications that are supported. Considering the ongoing effort towards providing convenient hardware platforms in industry, this paper focuses on design issu... | [
2988
] | Validation |
839 | 1 | Feature Subset Selection by Bayesian networks based optimization In this paper we perform a comparison among FSS-EBNA, a randomized, populationbased and evolutionary algorithm, and two genetic and other two sequential search approaches in the well known Feature Subset Selection (FSS) problem. In FSS-EBNA, the FSS problem, stated as a search problem, uses the EBNA (Estimation of Bayesian Network Algorithm) search engine, an algorithm within the EDA (Estimation of Distribution Algorithm) approach. The EDA paradigm is born from the roots of the GA community in order to explicitly discover the relationships among the features of the problem and not disrupt them by genetic recombination operators. The EDA paradigm avoids the use of recombination operators and it guarantees the evolution of the population of solutions and the discovery of these relationships by the factorization of the probability distribution of best individuals in each generation of the search. In EBNA, this factorization is carried out by a Bayesian network induced by a chea... | [
428,
999,
1516,
2641
] | Train |
840 | 1 | Fuzzy Concepts and Formal Methods: A Fuzzy Logic Toolkit for Z It has been recognised that formal methods are useful as a modelling tool in requirements engineering. Specification languages such as Z permit the precise and unambiguous modelling of system properties and behaviour. However some system problems, particularly those drawn from the IS problem domain, may be difficult to model in crisp or precise terms. It may also be desirable that formal modelling should commence as early as possible, even when our understanding of parts of the problem domain is only approximate. This paper suggests fuzzy set theory as a possible representation scheme for this imprecision or approximation. We provide a summary of a toolkit that defines the operators, measures and modifiers necessary for the manipulation of fuzzy sets and relations. We also provide some examples of the laws which establishes an isomorphism between the extended notation presented here and conventional Z when applied to boolean sets and relations. | [
860
] | Validation |
841 | 5 | Choosing Good Distance Metrics and Local Planners for Probabilistic Roadmap Methods This paper presents a comparative evaluation of different distance metrics and local planners within the context of probabilistic roadmap methods for motion planning. Both C-space and Workspace distance metrics and local planners are considered. The study concentrates on cluttered three-dimensional Workspaces typical, e.g., of mechanical designs. Our results include recommendations for selecting appropriate combinations of distance metrics and local planners for use in motion planning methods, particularly probabilistic roadmap methods. Our study of distance metrics showed that the importance of the translational distance increased relative to the rotational distance as the environment become more crowded. We find that each local planner makes some connections than none of the others do — indicating that better connected roadmaps will be constructed using multiple local planners. We propose a new local planning method we call rotate-at-s that outperforms the common straight-line in C-space method in crowded environments. | [
437,
693,
1003,
1936
] | Train |
842 | 1 | A Tutorial on Support Vector Machines for Pattern Recognition Abstract. The tutorial starts with an overview of the concepts of VC dimension and structural risk minimization. We then describe linear Support Vector Machines (SVMs) for separable and non-separable data, working through a non-trivial example in detail. We describe a mechanical analogy, and discuss when SVM solutions are unique and when they are global. We describe how support vector training can be practically implemented, and discuss in detail the kernel mapping technique which is used to construct SVM solutions which are nonlinear in the data. We show how Support Vector machines can have very large (even infinite) VC dimension by computing the VC dimension for homogeneous polynomial and Gaussian radial basis function kernels. While very high VC dimension would normally bode ill for generalization performance, and while at present there exists no theory which shows that good generalization performance is guaranteed for SVMs, there are several arguments which support the observed high accuracy of SVMs, which we review. Results of some experiments which were inspired by these arguments are also presented. We give numerous examples and proofs of most of the key theorems. There is new material, and I hope that the reader will find that even old material is cast in a fresh light. | [
1122
] | Train |
843 | 3 | PicASHOW: Pictorial Authority Search by Hyperlinks on the Web We describe PicASHOW, a fully automated WWW image retrieval system that is based on several link-structure analyzing algorithms. Our basic premise is that a page # displays (or links to) an image when the author of # considers the image to be of value to the viewers of the page. Wethus extend some well known link-based WWW #### ######### schemes to the context of image retrieval. PicASHOW's analysis of the link structure enables it to retrieve relevant images even when those are stored in les with meaningless names. The same analysis also allows it to identify ##### ########## and ##### ####. We dene these as Web pages that are rich in relevant images, or from which many images are readily accessible. PicASHOW requires no image analysis whatsoever and no creation of taxonomies for pre-classication of the Web's images. It can be implemented by standard WWW search engines with reasonable overhead, in terms of both computations and storage, and with no change to user query formats. It can thus be used to easily add image retrieving capabilities to standard search engines. Our results demonstrate that PicASHOW, while relying almost exclusively on link analysis, compares well with dedicated WWW image retrieval systems. We conclude that link analysis, a bona-de eective technique for Web page search, can improve the performance of Web image retrieval, as well as extend its denition to include the retrieval of image hubs and containers. Keywords Image Retrieval; Link Structure Analysis; Hubs and Authorities; Image Hubs. 1. | [
1201,
2459,
2503
] | Test |
844 | 3 | Boosting Interval-Based Literals: Variable Length and Early Classification In previous works, a system for supervised time series classification has been presented. It is based on boosting very simple classifiers: only one literal. The used predicates are based on temporal intervals. There are two types of predicates: i) relative predicates, such as "increases" and "stays", and ii) region predicates, such as "always" and "sometime", which operate ver regions in the domain of the variable. | [
1706
] | Train |
845 | 0 | Autonomy-Based Multi-Agent Systems: Statistical Issues This paper describes an autonomy-based multiagent system and its application to simulations in the framework of collective robotics. Experimental results measured for a particular task, namely object regrouping, have two outcomes. First, they show that a form of implicit cooperation takes place in the system despite the absence of explicit cooperation protocol, but just in virtue of the design of the system. Second, they indicate that it is possible to effect the way the task is achieved by appropriately introducing obstacles that modify the environment of the agents. A preliminary partial characterization of the concept of emergence is proposed and its illustration to our experimental system is given. Introduction Our work fits in the framework of Bottom-Up Artificial Intelligence (Brooks 1986), (Brooks 1991) and more particularly, in that of Autonomous Agents (Pfeifer 1995). We are concerned with collective phenomena and their issues and more precisely, the way to carry out solution... | [
1628
] | Train |
846 | 2 | Feature Subset Selection in Text-Learning This paper describes several known and some new methods for feature subset selection on large text data. Experimental comparison given on real-world data collected from Web users shows that characteristics of the problem domain and machine learning algorithm should be considered when feature scoring measure is selected. Our problem domain consists of hyperlinks given in a form of small-documents represented with word vectors. In our learning experiments naive Bayesian classifier was used on text data. The best performance was achieved by the feature selection methods based on the feature scoring measure called Odds ratio that is known from information retrieval. | [
830,
1002,
1440
] | Validation |
847 | 2 | Criterion Functions for Document Clustering: Experiments and Analysis In recent years, we have witnessed a tremendous growth in the volume of text documents available on the Internet, digital libraries, news sources, and company-wide intranets. This has led to an increased interest in developing methods that can help users to effectively navigate, summarize, and organize this information with the ultimate goal of helping them to find what they are looking for. Fast and high-quality document clustering algorithms play an important role towards this goal as they have been shown to provide both an intuitive navigation/browsing mechanism by organizing large amounts of information into a small number of meaningful clusters as well as to greatly improve the retrieval performance either via cluster-driven dimensionality reduction, term-weighting, or query expansion. This ever-increasing importance of document clustering and the expanded range of its applications led to the development of a number of new and novel algorithms with different complexity-quality trade-offs. Among them, a class of clustering algorithms that have relatively low computational requirements are those that treat the clustering problem as an optimization process which seeks to maximize or minimize a particular clustering criterion function defined over the entire clustering solution. | [
495,
1234,
1405,
2179,
2324
] | Train |
848 | 2 | Focused Crawls, Tunneling, and Digital Libraries Crawling the Web to build collections of documents related to pre-specified topics became an active area of research during the late 1990's after crawler technology was developed for the benefit of search engines. Now, Web crawling is being seriously considered as an important strategy for building large scale digital libraries. This paper considers some of the crawl technologies that might be exploited for collection building. For example, to make such collection-building crawls more effective, focused crawling was developed, in which the goal was to make a "best-first" crawl of the Web. We are using powerful crawler software to implement a focused crawl but use tunneling to overcome some of the limitations of a pure best-first approach. Tunneling has been described by others as not only prioritizing links from pages according to the page's relevance score, but also estimating the value of each link and prioritizing on that as well. We add to this mix by devising a tunneling focused crawling strategy which evaluates the current crawl direction on the fly to determine when to terminate a tunneling activity. Results indicate that a combination of focused crawling and tunneling could be an e#ective tool for building digital libraries. | [
471,
507,
608,
630,
1547,
1987,
2503,
2699,
2705
] | Test |
849 | 4 | Mining Usability Information from Log Files: AMulti-Pronged Approach rooms is configurable by its occupants in how they organize various tools housing their data, documents, and graphics. The TW system provides for synchronous and asynchronous user interactions, but importantly these interactions are in the context of relevant data. The work in this experiment was characterized by several full group meetings (for planning and coordination) interspersed with periods of individual activity (asychronous work) and smaller coordination meetings of two or three team members around the "hand-off" of output from a task used as input for another task. Collected Data The native version of TW produces a server-based log-file that contains information about the identity of users entering the distributed application, the identity of the rooms through which users navigate, file uploads, and message passing between users. This set of interactions was deemed too rudimentary for capturing the type of data needed for usability analysis. Since the source co | [
99,
1362
] | Train |
850 | 1 | Learning and Evaluating Visual Features for Pose Estimation We present a method for learning a set of visual landmarks which are useful for pose estimation. The landmark learning mechanism is designed to be applicable to a wide range of environments, and generalized for different approaches to computing a pose estimate. Initially, each landmark is detected as a local extremum of a measure of distinctiveness and represented by a principal components encoding which is exploited for matching. Attributes of the observed landmarks can be parameterized using a generic parameterization method and then evaluated in terms of their utility for pose estimation. We present experimental evidence that demonstrates the utility of the method. 1 Introduction In this paper, we develop an approach to visionbased robot localization by learning a set of imagedomain landmarks in the robot's environment. The landmarks are learned from a representative set of images obtained during an initial exploration of the environment. No a priori assumptions are made about the... | [
540,
548,
1404
] | Test |
851 | 0 | Real-Time Scheduling for Distributed Agents Providing an environment for a software agent to execute is very similar to building an operating system for the execution of general purpose applications. In the same fashion that an operating system provides a set of services for the execution of a user request, an agent framework provides a similar set of services for the execution of agent actions. Such services include the ability to communicate with other agents, maintaining the current state of an executing agent, and selecting an execution path from a set of possible execution paths. The particular focus of this paper is the study of Soft Real-Time agentscheduling in the context of a framework for the execution of intelligent software agents; acharacterization of agent performance; and developmentofanenvironment for testing and comparing the performance of agent activities. The agent architecture used for this study, DECAF #Distributed Environment Centered AgentFramework#, is a software toolkit for the rapid d... | [
2364
] | Test |
852 | 0 | Constraints and Agents in MADEsmart As part of the DARPA Rapid Design Exploration and Optimization (RaDEO) program, Boeing, Philadelphia, is involved in an on-going concurrent design engineering research project called MADEsmart which seeks to partially automate the Integrated Product Team (IPT) concept used by Boeing for organizing the design engineering process, with the aid of intelligent agent technology. Although currently only in an early stage of development, the project is expected to crucially employ a constraint-centered System Design Management Agent developed by the University of Toronto's IE Department in conjunction with Boeing. The SDMA will use the constraint-based Toronto Ontologies for a Virtual Enterprise (TOVE) ontologies, and its domain theories for design engineering and dependent underlying theories, phrased as KIF/Ontolingua assertions in an axiomatic system running in the constraint logic system ECLiPSe, as its primary knowledge resource to monitor an ongoing design project, offering resource-all... | [
1107
] | Test |
853 | 3 | Scalable Algorithms for Large-Scale Temporal Aggregation The ability to model time-varying natures is essential to many database applications such as data warehousing and mining. However, the temporal aspects provide many unique characteristics and challenges for query processing and optimization. Among the challenges is computing temporal aggregates, which is complicated by having to compute temporal grouping. In this paper, we introduce a variety of temporal aggregation algorithms that overcome major drawbacks of previous work. First, for small-scale aggregations, both the worst-case and average-case processing time have been improved significantly. Second, for large-scale aggregations, the proposed algorithms can deal with a database that is substantially larger than the size of available memory. Third, the parallel algorithm designed on a shared-nothing architecture achieves scalable performance by delivering nearly linear scale-up and speed-up. The contributions made in this paper are particularly important because the rate of increase ... | [
15,
1661
] | Train |
854 | 4 | Feasibility Discussion of a Collaborative Virtual Environment - FINDING ALTERNATIVE WAYS FOR UNIVERSITY MEMBERS INTERACTION This paper discusses the potential impact and roadmap for the creation of a Collaborative Virtual Environment where all university members can interact in novel ways. Some actual NetLab figures are presented to justify this evolution as feasible. A related project that uses the potential created by the "laptops for all" action is a virtual incubator to simulate entrepreneurship bias is presented. | [
1933
] | Train |
855 | 2 | Automatic Multi-Lingual Information Extraction Information Extraction (IE) is a burgeoning technique because of the explosion of internet. So far, most of the IE systems are focusing on English text; and most of them are in the supervised learning framework, which requires large amount of human labor; and most of them can only work in narrow domain, which is domain dependent. These systems are difficult to be ported to other languages, other domains because of these inherent shortcomings. Currently, besides western languages like English, there are many other Asian languages which are much di erent from English. In English, words are delimited by white-spaces so computer can easily tokenize the input text string. In many languages like Chinese, Japanese, Thai and Korea, they do not have word boundaries between words. This poses a difficult problem for the information extraction for those languages. In this thesis, we intend to implement a self-contained, language independent automatic IE system. The system is automatic because we are using a unsupervised learning framework in which no labeled data is required for training or a semi-supervised learning framework in which small amount of labeled data and large amount of unlabeled data are used. Specifically, we deal with Chinese and English languages name entity recognition and entity relation extraction, but the system can be easily extended to any other languages and other tasks. We implement an unsupervised Chinese word segmenter, a Chinese POS tagger, and we extend maximum entropy models to incorporate unlabeled data for general information extraction. | [
875,
988,
1159,
1280,
1290,
1366,
1386,
2176,
2379,
2443
] | Test |
856 | 3 | The Vagabond Parallel Temporal Object-Oriented Database System: Versatile Support for Future Applications . In this paper, we discuss features that future database systems should support to deliver the required functionality and performance to future applications. The most important features are efficient support for: 1) large objects, 2) isochronous delivery of data, 3) queries on large data sets, 4) full text indexing, 5) multidimensional data, 6) sparse data, and 7) temporal data and versioning. To efficiently support these features in one integrated system, a new database architecture is needed. We describe an architecture suitable for this purpose, the Vagabond Parallel Temporal Object-Oriented Database system. We also describe techniques we have developed to avoid some possible bottlenecks in a system based on this new architecture. 1 Introduction The recent years have brought computers into almost every office, and this availability of powerful computers, connected in global networks, has made it possible to utilize powerful data management systems in new application areas.... | [
1803
] | Train |
857 | 3 | Simplifying Data Access: The Energy Data Collection (EDC) Project The massive amount of statistical and text data available from government agencies has created a set of daunting challenges to both research and analysis communities. These problems include heterogeneity, size, distribution, and control of terminology. At the Digital Government Research Center we are investigating solutions to these key problems. In this paper we focus on (1) ontological mappings for terminology standardization, (2) data integration across data bases with high speed query processing, and (3) interfaces for query input and presentation of results. This collaboration between researchers from Columbia University and the Information Sciences Institute of the University of Southern California employs technology developed at both locations, in particular the SENSUS ontology, the SIMS multi-database access planner, the LKB automated dictionary and terminology analysis system, and others. The pilot application targets gasoline data from the Bureau of Labor Statistics, the Energy Information Administration of the Department of Energy, the Census Bureau, and other government agencies. 1 | [
1127,
2205
] | Test |
858 | 1 | Cooperative Learning over Composite Search Spaces: Experiences with a Multi-agent Design System We suggest the use of two learning techniques --- short term and long term --- to enhance search efficiency in a multi-agent design system by letting the agents learn about non-local requirements on the local search process. The first technique allows an agent to accumulate and apply constraining information about global problem solving, gathered as a result of agent communication, to further problem solving within the same problem instance. The second technique is used to classify problem instances and appropriately index and retrieve constraining information to apply to new problem instances. These techniques will be presented within the context of a multi-agent parametricdesign application called STEAM. We show that learning conclusively improves solution quality and processingtime results. Introduction In this article, we study machine-learning techniques that can be applied within multi-agent systems (MAS) to improve solution quality and processing-time results. A ubiquitous prob... | [
286,
1518,
2877
] | Train |
859 | 1 | Beyond Euclidean Eigenspaces: Bayesian Matching for Visual Recognition We propose a novel technique for direct visual matching of images for the purposes of face recognition and database search. Speci#cally,we argue in favor of a probabilistic measure of similarity, in contrast to simpler methods which are based on standard Euclidean L2 norms #e.g., template matching# or subspace-restricted norms #e.g., eigenspace matching#. The proposed similarity measure is based on a Bayesian analysis of image di#erences: we model twomutually exclusive classes of variation between two facial images: intra-personal #variations in appearance of the same individual, due to di#erent expressions or lighting# and extra-personal #variations in appearance due to a di#erence in identity#. The high-dimensional probability density functions for each respective class are then obtained from training data using an eigenspace density estimation technique and subsequently used to compute a similarity measure based on the a posteriori probability of membership in the intra-personal class, which is used to rank matches in the database. The performance advantage of this probabilistic matching technique over standard Euclidean nearest-neighbor eigenspace matching is demonstrated using results from ARPA's 1996 #FERET" face recognition competition, in which this algorithm was found to be the top performer. | [
3058
] | Validation |
860 | 5 | Fuzzy Concepts and Formal Methods: Some Illustrative Examples It has been recognised that formal methods are useful as a modelling tool in requirements engineering. Specification languages such as Z permit the precise and unambiguous modelling of system properties and behaviour. However some system problems, particularly those drawn from the IS problem domain, may be difficult to model in crisp or precise terms. It may also be desirable that formal modelling should commence as early as possible, even when our understanding of parts of the problem domain is only approximate. This paper identifies the problem types of interest and argues that they are characterised by uncertainty and imprecision. It suggests fuzzy set theory as a useful formalism for modelling aspects of this imprecision. The paper illustrates how a fuzzy logic tooklit for Z can be applied to such problem domains. Several examples are presented illustrating the representation of imprecise concepts as fuzzy sets and relations, soft pre- conditions and system requirements as a series of linguistically quantified propositions. 1 | [
840
] | Train |
861 | 3 | Constructing and Transforming CBR Implementations: Techniques for Corporate Memory Management Achieving widespread case-based reasoning support for corporate memories will require the flexibility to integrate implementations with existing organizational resources and infrastructure. Case-based reasoning implementations as currently constructed tend to fall into three categories, characterized by implementation constraints: task-based (task constraints alone), enterprise (integrating databases), and web-based (integrating web representations) . These implementation types represent the possible targets in constructing corporate memory systems, and it is important to understand the strengths of each, how they are built, and how one may be constructed by transforming another. This paper describes a framework that relates the three types of CBR implementation, discusses their typical strengths and weaknesses, and describes practical strategies for building corporate CBR memories to meet new requirements by transforming and synthesizing existing resources. 1 Introduction Constructi... | [
2592
] | Train |
862 | 3 | Using Database Optimization Techniques for Nonmonotonic Reasoning In this paper a program rewriting technique for disjunctive datalog is proposed, which descends from query optimization techniques in relational algebra and reduces the size of the ground instantiation of a program in many cases. As a consequence, the time for generating the ground instantiation and the time for subsequent operations on the ground program is shortened. A part of this technique has already been implemented as a preprocessing step in the Disjunctive Deductive Database System dlv. Using a recently published application, we show that a signicant reduction of the ground program size and a tremendous overall speedup can be achieved. 1 Introduction dlv is a Disjunctive Deductive Database System implementing the Consistent Answer Sets Semantics. Like similar systems in the area of non-monotonic reasoning [6], the kernel modules of dlv operate on a ground instantiation of the input program, i.e., a program that does not contain any variables, but is (semantically) equivalent t... | [
1919
] | Train |
863 | 3 | Similarity Query Processing in Image Databases CHITRA is a prototype CBIR system we are building. It uses a four layer data model we have developed, and enables retrieval based on high level concepts, such as "retrieve images of MOUNTAINS", and "retrieve images of MOUNTAINS and SUNSET". This paper deals with some issues about query processing encountered in the implementation of the system. The contributions of this paper can be summarized in terms of processing the following four example queries (I 1 , I 2 , ..,I k are images). Q 1 : "retrieve images similar to I 1 based on color". Q 2 : "retrieve images similar to I 1 based on color AND texture". Q 3 : "retrieve images similar to I 1 ; I 2 ; ::; I k based on color". Q 4 : "retrieve images similar to I 1 ; I 2 ; ::; I k based on color AND texture". First a brief review of basic CBIR query processing literature is provided (processing of Q 1 ). Processing of Q 2 involves efficient evaluation of combining functions, a problem that has attracted research attention in recent time... | [
1858
] | Test |
864 | 5 | The CMUnited-97 Robotic Soccer Team: Perception and Multiagent Control Robotic soccer is a challenging research domain which involves multiple agents that need to collaborate in an adversarial environment to achieve specificobjectives. In this paper, we describe CMUnited, the team of small robotic agents that we developed to enter the RoboCup-97 competition. We designed and built the robotic agents, devised the appropriate vision algorithm, and developed and implemented algorithms for strategic collaboration between the robots in an uncertain and dynamic environment. The robots can organize themselves in formations, hold specificroles, and pursue their goals. In game situations, they have demonstrated their collaborative behaviors on multiple occasions. The robots can also switch roles to maximize the overall performance of the team. We present an overview of the vision processing algorithm which successfully tracks multiple moving objects and predicts trajectories. The paper then focusses on the agent behaviors ranging from low-level individual behaviors to coordinated, strategic team behaviors. | [
346,
962,
1630,
2264,
2492,
3173
] | Validation |
865 | 2 | Effective Site Finding using Link Anchor Information Link-based ranking methods have been described in the literature and applied in commercial Web search engines. However, according to recent TREC experiments, they are no better than traditional content-based methods. We conduct a different type of experiment, in which the task is to find the main entry point of a specific Web site. In our experiments, ranking based on link anchor text is twice as effective as ranking based on document content, even though both methods used the same BM25 formula. We obtained these results using two sets of 100 queries on a 18.5 million docu- ment set and another set of 100 on a 0.4 million document set. This site finding effectiveness begins to explain why many search engines have adopted link methods. It also opens a rich new area for effectiveness improvement, where traditional methods fail. | [
471,
1838,
2503,
2705
] | Test |
866 | 2 | Improving Short-Text Classification Using Unlabeled Background Knowledge to Assess Document Similarity We describe a method for improving the classification of short text strings using a combination of labeled training data plus a secondary corpus of unlabeled but related longer documents. We show that such unlabeled background knowledge can greatly decrease error rates, particularly if the number of examples or the size of the strings in the training set is small. This is particularly useful when labeling text is a labor-intensive job and when there is a large amount of information available about a particular problem on the World Wide Web. Our approach views the task as one of information integration using WHIRL, a tool that combines database functionalities with techniques from the information-retrieval literature. 1. Introduction The task of classifying textual data that has been culled from sites on the World Wide Web is both difficult and intensively studied (Cohen & Hirsh, 1998; Joachims, 1998; Nigam et al., 1999). Applications of various machine learning techniqu... | [
1386,
1478,
1843,
2864
] | Train |
867 | 0 | Composable Agents for Patient Flow Control - Preliminary Concepts In this article we describe our research efforts in coping with a trade-off that can be often found in the control and optimization of todays business processes. Though centralized control may achieve better results in controlling the system behavior, there are usually social, technical and security constraints on applying centralized control. Distributed control on the other hand may cope with these constraints but also entails suboptimal results and communicational overhead. Our concept of composable agents tries to allow a dynamic and fluent transition between globalization and localization in business process control by adapting to the current real-world system structure. We are currently evaluating this concept in the framework of a patient flow control project at Charit'e Berlin. Todays applications of information technology face at least two major aspects of business settings. The first aspect is the partially or fully automated execution of complex business processes. This enfo... | [
49,
2364
] | Train |
868 | 2 | Building Infrastructures for Digital Libraries Digital Libraries today are often monolithic systems. In the future, they will dissolve into collections of electronic services. The challenge will be to provide an infrastructure that supports the user in dealing with this multitude of services. Such an infrastructure should offer integrated access to the combined contents of multiple services, it should provide active dissemination of new contents, and it needs to support the users in locating and combining the services most suitable to their needs. In the Global Info program Infrastructures for Digital Libraries components of such an infrastructure are being developed. The federated query service DEMETRIOS and the alerting service HERMES are both integration services that combine underlying services, i.e., heterogenous information sources. The GIBRALTAR portal provides a meta-service that supports the user in locating and applying various Digital Library services. 1 | [
35
] | Train |
869 | 3 | A Query Calculus for Spatio-Temporal Object Databases The development of any comprehensive proposal for spatio-temporal databases involves significant extensions to many aspects of a non-spatio-temporal architecture. One aspect that has received less attention than most is the development of a query calculus that can be used to provide a semantics for spatio-temporal queries and underpin an effective query optimization and evaluation framework. In this paper, we show how a query calculus for spatiotemporal object databases that builds upon the monoid calculus proposed by Fegaras and Maier for ODMG-compliant database systems can be developed. The paper shows how an extension of the ODMG type system with spatial and temporal types can be accommodated into the monoid approach. It uses several queries over historical (possibly spatial) data to illustrate how, by mapping them into monoid comprehensions, the way is open for the application of a logical optimizer based on the normalization algorithm proposed by Fegaras and Maier. | [
131,
460,
653,
1068,
1101,
1635
] | Train |
870 | 0 | Making Complex Articulated Agents Dance - An analysis of control methods drawn from robotics, animation, and biology . We discuss the tradeoffs involved in control of complex articulated agents, and present three implemented controllers for a complex task: a physically-based humanoid torso dancing the Macarena. The three controllers are drawn from animation, biological models, and robotics, and illustrate the issues of joint-space vs. Cartesian space task specification and implementation. We evaluate the controllers along several qualitative and quantitative dimensions, considering naturalness of movement and controller flexibility. Finally, we propose a general combination approach to control, aimed at utilizing the strengths of each alternative within a general framework for addressing complex motor control of articulated agents. Key words: articulated agent control, motor control, robotics, animation 1. Introduction Control of humanoid agents, dynamically simulated or physical, is an extremely difficult problem due to the high dimensionality of the control space, i.e., the many degrees of freed... | [
401
] | Validation |
871 | 3 | Segment-Based Approach for Subsequence Searches in Sequence Databases This paper investigates the subsequence searching problem under time warping in sequence databases. Time warping enables to find sequences with similar changing patterns even when they are of different lengths. Our work is motivated by the observation that subsequence searches slow down quadratically as the total length of data sequences increases. To resolve this problem, we propose the SegmentBased Approach for Subsequence Searches (SBASS), which modifies the similarity measure from time warping to piecewise time warping and limits the number of possible subsequences to be compared with a query sequence. For efficient retrieval of similar subsequences without false dismissal 1 , we extract feature vectors from all data segments exploiting their monotonically changing properties, and build a multi-dimensional index such as R-tree or R - tree. Using this index, queries are processed with four steps: 1) index filtering, 2) feature filtering, 3) successor filtering, and 4) post-proce... | [] | Validation |
872 | 3 | Abduction in Logic Programming This paper is a survey and critical overview of recent work on the extension of Logic Programming to perform Abductive Reasoning (Abductive Logic Programming). We outline the general framework of Abduction and its applications to Knowledge Assimilation and Default Reasoning; and we introduce an argumentation-theoretic approach to the use of abduction as an interpretation for Negation as Failure. We also analyse the links between Abduction and the extension of Logic Programming obtained by adding a form of explicit negation. Finally we discuss the relation between Abduction and Truth Maintenance. 1 Introduction This paper is a survey and analysis of work on the extension of logic programming to perform abductive reasoning. The purpose of the paper is to provide a critical overview of some of the main research results, in order to develop a common framework for evaluating these results, to identify the main unresolved problems, and to indicate directions for future work. The emphasis i... | [
441,
1185,
1325,
1625,
1760,
1828,
1833,
2678
] | Train |
873 | 2 | Automatic Pattern Acquisition for Japanese Information Extraction One of the central issues for information extraction is the cost of customization from one scenario to another. Research on the automated acquisition of patterns is important for portability and scalability. In this paper, we introduce Tree-Based Pattern representation where a pattern is denoted as a path in the dependency tree of a sentence. We outline the procedure to acquire Tree-Based Patterns in Japanese from un-annotated text. The system extracts the relevant sentences from the training data based on TF/IDF scoring and the common paths in the parse tree of relevant sentences are taken as extracted patterns. Keywords Information Extraction, Pattern Acquisition 1. | [
1546
] | Train |
874 | 3 | Towards a Model for Spatio-Temporal Schema Selection Schema versioning provides a mechanism for handling change in the structure of database systems and has been investigated widely, both in the context of static and temporal databases. With the growing interest in spatial and spatio-temporal data as well as the mechanisms for holding such data, the spatial context within which data is formatted also becomes an issue. This paper presents a generalised model that accommodates schema versioning within static, temporal, spatial and spatio-temporal relational and object-oriented databases. | [
2935,
3100
] | Validation |
875 | 2 | Scenario Customization for Information Extraction Information Extraction (IE) is an emerging NLP technology, whose function is to process unstructured, natural language text, to locate specific pieces of information, or facts, in the text, and to use these facts to fill a database. IE systems today are commonly based on pattern matching. The core IE engine uses a cascade of sets of patterns of increasing linguistic complexity. Each pattern consists of a regular expression and an associated mapping from syntactic to logical form. The pattern sets are customized for each new topic, as defined by the set of facts to be extracted. Construction of a pattern base for a new topic is recognized as a time-consuming and expensive process---a principal roadblock to wider use of IE technology in the large. An e#ective pattern base must be precise and have wide coverage. This thesis addresses the portability probl... | [
714,
855,
2068,
2443
] | Train |
876 | 3 | Integrity Constraints: Semantics and Applications this paper.) Finally, by having ICs expressed in logic, one can use deduction and logic as one's basic and natural tools for handling, and reasoning over, database specifications (facts, rules, queries, and ICs). This applies for relational databases, with or without rules, and can be applied to object-oriented and object-relational databases as well. There is a broad body of work on logic and relational databases, and a general consensus on what databases (facts and rules) and queries mean. However, there is less work on the meaning of integrity constraints, and certainly no consensus. What is meant by an IC can differ widely. For instance, one may define that ICs must be consistent with the database, or define that they must be provable statements, deducible from the database. (This distinction is discussed in more detail in Section 3.2.) Another view is that ICs really represent meta-knowledge---knowledge about the database itself---and should, perhaps, be written in a nonclassical logic. The general situation becomes more complex when we permit databases to contain indefinite or disjunctive information or to use negation, as with IC 2 above. Subtle but profound differences in meaning can arise due to different interpretations of ICs. In many systems, the semantics for ICs is never made clear; at times, one interpretation seems intended, while at other times, another interpretation is evident. This ambiguity is dangerous, and could allow a database to become corrupt in unanticipated ways. In this chapter, we introduce integrity constraints in a logical framework and overview the various work that has been done on ICs in this context. Section 2 presents the basic definitions for databases and integrity constraints for the chapter. We consider various semantics of ICs,... | [
1007,
1408,
1537,
2347,
2821
] | Train |
877 | 3 | Algorithms for Temporal Query Operators in XML Databases The contents of an XML database or XML/Web data warehouse is seldom static. New documents are created, documents are deleted, and more important: documents are updated. In many cases, we want to be able to search in historical versions, retrieve documents valid at a certain time, query changes to documents, etc. This can be supported by extending the system with temporal database features. In this paper we describe the new query operators needed in order to support an XML query language which supports temporal operations. We also describe the algorithms which can make efficient implementation of these query operators possible. Keywords: XML, temporal databases, query processing 1 | [
416,
1660,
2232,
2524,
2632
] | Train |
878 | 2 | Probabilistic Hierarchical Clustering with Labeled and Unlabeled Data . This paper presents hierarchical probabilistic clustering methods for unsupervised and supervised learning in datamining applications, where supervised learning is performed using both labeled and unlabeled examples. The probabilistic clustering is based on the previously suggested Generalizable Gaussian Mixture model and is extended using a modified Expectation Maximization procedure for learning with both unlabeled and labeled examples. The proposed hierarchical scheme is agglomerative and based on probabilistic similarity measures. Here, we compare a L 2 dissimilarity measure, error confusion similarity, and accumulated posterior cluster probability measure. The unsupervised and supervised schemes are successfully tested on artificially data and for e-mails segmentation. 1 | [
890,
1386,
2055,
2468,
2889
] | Train |
879 | 2 | Theme-Based Retrieval of Web News Efficient information retrieval of highly dynamic information, such as Web news, is a complex task. As a result, search and retrieval environments for continuously updated news from other sources than the largest media conglomerates are almost absent on the Internet. Global search engines do not index or classify news information from smaller network communities. To address this problem, I developed NewsSearch, a news information management environment designed to improve retrieval efficiency of online news for the smaller networked communities. NewsSearch search achieves its goal through a combination of techniques: . Multiple indexing queues, defining multiple gathering schedules, to deal with different publication periodicities. . Information Retrieval techniques to news, in order to classify them into a pre-defined set of themes. . Support Vector Machines, which proved to be a fast and reliable classification technique. NewsSearch proved to be a scalable solution with acceptable storage needs even while managing a fairly large collection of daily publications. A combination of fine tuning of training strategies, noise filtering of Web news documents and multiple classifications, enable NewsSearch to achieve a classification accuracy of 95%. ACKNOWLEDGEMENTS This work was supported in part by the PRAXIS project ARIADNE (Pblico Digital -- Praxis XXI, Medida 3.1b) and project SAGRES (Praxis/P/TIT/1676/95). TABLE OF CONTENTS CHAPTER I | [
1120,
2275,
2464
] | Train |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.