Search

The increasing number of satellite missions for Earth exploration provides huge amounts of data. These data are of wide diversity regarding the images characteristics, thus new techniques and tools need to be developed to accommodate the extraction of meaningful information. This paper presents An Earth Observation Spatio-Temporal Data Mining System (HyperMINE), which integrates fast and complex query methods in order to generate SITS (Satellite Image Time Series) regarded as a data hypercube. The multidimensional data, considering geographical space, time, band, and satellite/sensor are used as input for information mining algorithms. The system is built on a modular multilayer architecture that allows effective processing of various sources of data.

In this paper, we deal with the integration of multiple sources of information such as Earth observation (EO) synthetic aperture radar (SAR) images and their metadata, semantic descriptors of the image content, as well as other publicly available geospatial data sources expressed as linked open data for posing complex queries in order to support geospatial data analytics. Our approach lays the foundations for the development of richer tools and applications that focus on EO image analytics using ontologies and linked open data. We introduce a system architecture where a common satellite image product is transformed from its initial format into to actionable intelligence information, which includes image descriptors, metadata, image tiles, and semantic labels resulting in an EO-data model. We also create a SAR image ontology based on our EO-data model and a two-level taxonomy classification scheme of the image content. We demonstrate our approach by linking high-resolution TerraSAR-X images with information from CORINE Land Cover (CLC), Urban Atlas (UA), GeoNames, and OpenStreetMap (OSM), which are represented in the standard triple model of the resource description frameworks (RDFs).

This paper illustrates the usefulness of natural language processing (NLP) tools for computer assisted language learning (CALL) through the presentation of three NLP tools integrated within a CALL software for French. These tools are (i) a sentence structure viewer; (ii) an error diagnosis system; and (iii) a conjugation tool. The sentence structure viewer helps language learners grasp the structure of a sentence, by providing lexical and grammatical information. This information is derived from a deep syntactic analysis. Two different outputs are presented. The error diagnosis system is composed of a spell checker, a grammar checker, and a coherence checker. The spell checker makes use of alpha-codes, phonological reinterpretation, and some ad hoc rules to provide correction proposals. The grammar checker employs constraint relaxation and phonological reinterpretation as diagnosis techniques. The coherence checker compares the underlying "semantic" structures of a stored answer and of the learners' input to detect semantic discrepancies. The conjugation tool is a resource with enhanced capabilities when put on an electronic format, enabling searches from inflected and ambiguous verb forms.

Finding maneuvers for parameter estimation in flight data records is a laborious task and traditionally performed post-flight on ground. Two different data parser approaches to automatically detect these maneuvers in flight are presented. Both methods search the control input signals for significant changes that correspond to test maneuvers. The first algorithm is based on the signal time derivative of the input signal whereas the second method uses a fast orthogonal wavelet transform. Both algorithms are tested with flight test data recorded with the DLR research aircraft ATTAS. Performance results are compared and potential problems when applying the parsers to other data are discussed. Results indicate that both methods are applicable in an online parameter estimation tool. The intention of the work in this paper is to develop an algorithm with a high level of automation for in-flight use, but both approaches could also be applied to offline flight data mining problems.

This paper discusses pivotal theoretical and methodological problems of historical corpus linguistics. In two case studies from Swedish language history, the development of the epistemic adverb kanske and the group genitive respectively, it illustrates how the use of qualitative method in addition to corpus investigation can contribute to understanding grammatical change.

Hyponyms of place in Polish and American wordnets The aim of the paper is to compare the sets of hyponyms of place in Polish and American wordnets. Polish “Słowosieć” and American WordNet are thesauri which define the meaning of an individual lexeme by the lexical-semantic relations existing between this and other lexemes. The researched material consists of several thousand of synsets (sets of lexemes referring to a common concept) which have been linked by interlingual, lexical-semantic relations to the synsets of English lexemes in a process of mapping of a fragment of the Polish wordnet to the American database. An analysis of particular cases illustrates the basic issues concerning the procedure of mapping, first of all the problem of choosing a foreign synonym for the original word. It also shows the incompatibility of the corresponding fragments of the Polish and English lexical systems, which is usually not reflected in dictionaries and other glottodidactical publications. Differences between the structures of semantic trees of the two languages result partly from different assumptions made by the authors of both wordnets, and partly from language-specific factors, such as different word-formation systems or different ways in which some objects from the real world are conceptualized.

Dimensionality reduction is the most widely used approach for extracting the most informative low-dimensional features from highdimensional ones. During the last two decades, different techniques (linear and nonlinear) have been proposed by researchers in various fields. However, the main question is now how well a specific technique does this job. In this paper, we introduce a qualitative method to assess the quality of dimensionality reduction. In contrast to numerical assessment, we focus here on visual assessment. We visualize the Minimum Spanning Tree (MST) of neighborhood graphs of data before and after dimensionality reduction in an immersive 3D virtual environment. We employe a mixture of linear and nonlinear dimension reduction techniques to apply to both synthetic and real datasets. The visualization depicts the quality of each technique in term of preserving distances and neighborhoods. The results show that a specific dimension reduction technique exhibits different performance in dealing with different datasets.

We are dealing with large-scale high-dimensional image data sets requiring new approaches for data mining where visualization plays the main role. Dimension reduction (DR) techniques are widely used to visualize high-dimensional data. However, the information loss due to reducing the number of dimensions is the drawback of DRs. In this paper, we introduce a novel metric to assess the quality of DRs in terms of preserving the structure of data. We model the dimensionality reduction process as a communication channel model transferring data points from a high-dimensional space (input) to a lower one (output). In this model, a co-ranking matrix measures the degree of similarity between the input and the output. Mutual information (MI) and entropy defined over the co-ranking matrix measure the quality of the applied DR technique. We validate our method by reducing the dimension of SIFT and Weber descriptors extracted from Earth Observation (EO) optical images. In our experiments, Laplacian Eigenmaps (LE) and Stochastic Neighbor Embedding (SNE) act as DR techniques. The experimental results demonstrate that the DR technique with the largest MI and entropy preserves the structure of data better than the others.

While software building blocks and their interdependencies can be recovered from the source code using static fact extraction, behavior and communication paths at runtime are typically gathered from instrumented executions of the system. However, more often than not it is not possible to retrieve data from the running system- either due to a high effort for instrumentation, missing (hardware) infrastructure, or because of advanced communication mechanisms hidden by middle ware, frameworks or platforms. In this paper, we present an approach to semi-automatically reconstruct runtime components and connectors using source code analysis, pattern matching, and expert knowledge. We present two applications where we could recover runtime communication paths and component interactions despite the absence of runtime traces.

In this position paper we describe how technologies from the Web of Things domain could help to simplify and automate the interaction between robotic devices and their surroundings. Specifically, we discuss how Web patterns like Resource-oriented Architectures or Representational State Transfer together with semantic metadata could help to create environments where robots seamlessly interact with other devices in their vicinity: Using services provided by other devices and offering services themselves. One of the main goals in the Web of Things community is to furnish smart environments with sensors and actuators that offer self-described interfaces and are openly accessible from other devices -- Robots should be enabled to make use of this wealth of instrumentation of their surroundings!

Unlike traditional networked embedded systems, the Internet of Things interconnects heterogeneous devices from various manufacturers with diverse functionalities. To foster the emergence of novel applications, this vast infrastructure requires a common application layer. As a single global standard for all device types and application domains is impracticable, we propose an architecture where the infrastructure is agnostic of applications and application development is fully decoupled from the embedded domain. In our design, the application logic of devices is running on application servers, while thin servers embedded into devices export only their elementary functionality using REST resources. In this paper, we present our design goals and preliminary results of this approach, featuring the Californium (Cf) CoAP framework.

This paper is focused on modern approaches to machine learning, most of which are as yet used infrequently or not at all in chemoinformatics. Machine learning methods are characterized in terms of the "modes of statistical inference" and "modeling levels" nomenclature and by considering different facets of the modeling with respect to input/ouput matching, data types, models duality, and models inference. Particular attention is paid to new approaches and concepts that may provide efficient solutions of common problems in chemoinformatics: improvement of predictive performance of structure-property (activity) models, generation of structures possessing desirable properties, model applicability domain, modeling of properties with functional endpoints (e.g., phase diagrams and dose-response curves), and accounting for multiple molecular species (e.g., conformers or tautomers).

The application of modeling and simulation in the design, development and validation process of complex systems has significantly increased in the last decades. Creating high quality models is a time-consuming task. Particularly, if models should be shared and reused in future projects. In this paper, results of the project Simulation Model Library (SimMoLib) are presented that address the issues concerning the preservation of knowledge that lies within simulation and calculation models. SimMoLib provides modeling guidelines and best practices to help the developer to prepare models that can be reused in other contexts. Validation and verification of these models is covered by proposing a set of guidelines and two test frameworks as well as by promoting peer reviews of models by other experts. To archive, catalogue, and distribute models, a software framework is under development which supports the creation, management, retrieval, and utilization of models. In order to allow the collaborative editing of calculation and simulation models a simplified version control mechanism is established. SimMoLib is currently targeted to support the development of space systems. In the future it will be opened to other domains.

The article focuses on some possibilities of using the available materials on the interactive board, which can contribute to the development of the learning strategies and to the quality as well as to the efficiency in the process of vocabulary learning. The current findings on learning strategies and vocabulary learning (Oxford 1990, Nunan 1999, Nation 2001) are combined and compared to the usage of various materials of i-board in the classroom. A meaningful usage of i-board being our priority, we found more advantages in the use of i-board in comparison to the classical exercises on the paper. Our material is based on Piepho's and Häussermann's typology of exercises for improving vocabulary learning. The students' perception of the interactive board in classroom is included and is based on analysis of the results of an online survey of 58 students, who had been using interactive board in German language classes for two years. V članku se ukvarjamo z možnostmi uporabe interaktivne table, s katerimi lahko prispevamo h kakovostnejšemu in učinkovitejšemu učenju besedišča ter vplivamo na razvijanje učnih strategij za učenje besedišča. Naša izhodišča predstavljajo dognanja s področja učenja besedišča in učnih strategij (Oxford 1990, Nunan 1999, Nation 1999), ki jih poskušamo povezati z uporabo interaktivne table v razredu. Predstavljamo možnosti smiselne uporabe i-table, kjer smo ugotovili prednosti pred vadenjem in urjenjem na papirju, ter nekaj primerov vaj in nalog po tipologiji Häussermanna in Piephoja s poudarkom na usvajanju besedišča in razvijanju učnih strategij. Na podlagi analize rezultatov spletne ankete o uporabi interaktivne table in njenem vplivu na utrjevanje in preverjanje besedišča, ki je bila izvedena pri pouku nemščine v 4. letniku, predstavljamo učenčeva dojemanja uporabe i-table v razredu.

The question of regularity within morphological paradigms has been formerly addressed within approaches falling in the scope of Canonical Typology (Corbett 2003). The aim of this paper is to provide a means for assessing the notion of morphological canonicity through original measures developed within our new morphological framework PARSLI. In particular, we introduce original measures for non-canonical phenomena such as heteroclisis, deponency, defectiveness and overabundance. We introduce PARSLI a new model for inflectional morphology using an inferential-realisational approach (Matthews 1974; Zwicky 1985; Anderson 1992). Our model precisely provides a formal representation of the lexicon/grammar interface. It relies on a formal definition of a lexical entry and a complete formal apparatus for computing all relevant form realisation rules for each lexeme, including stem formation rules. Realisation rules themselves may be expressed through any suitable realisation-based formalism (e.g. PFM or Network Morphology). We introduce several formal innovations such as inflection zones, that constitute partitions of given inflection classes. They are in particular used in modelling heteroclisis. Vprašanja pravilnosti morfoloških paradigem so se že lotevali pristopi, ki sodijo v okvir kanonične tipologije (Corbett 2003). Cilj pričujočega članka je prispevati izvirne načine, ki bodo na podlagi meril, ki smo jih izdelali znotraj našega novega morfološkega modela PARSLI, omogočali ovrednotiti pojem morfološke kanoničnosti. Še posebej pozorno pa smo vpeljali nove načine merjenja nekanoničnih pojavov, kot smo npr. heterokliza, deponentnost, nezapolnjnenost, prenapolnjenost. V članku predstavljamo PARSLI, ki je nov oblikoslovni model, ki se opira na inferenčnouresničitveni pristop (Matthews 1974; Zwicky 1985; Anderson 1992). Naš model ponuja prav formalno predstavitev slovarsko-slovničnega vmesnika. Temelji na formalni definiciji leksikalne iztočnice in popolnem formalnem aparatu, ki omogoča izpeljavo vseh relevantnih oblikoslovnih uresničitvenih pravil za vsak leksem, kamor sodijo tudi pravila oblikovanja osnove. Uresničitvena pravila lahko oblikujemo znotraj katerega koli ustreznega formalnega modela (na primer, teorija paradigmatskih funkcij ali morfologija mrež /ang. Network Morphology/). Vpeljemo vrsto formalnih novosti, na primer pregibna območja (ang. inflection zones), ki tvorijo dele posameznega pregibnega razreda. Posebej koristni so pri modeliranju heteroklize.

Traditional dictionaries provide lexicographic data of the same type in different ways; such unrestricted freedom might result in imprecise information. Some such cases are explained by an attempt to compile traditional dictionaries so that they are handy for their users. The paper discusses several non-systemic cases of providing forms of lexicographic items, their meanings and shades of meaning, derivatives, proper and common lexis. The data has been drawn from more than ten dictionaries of several languages. The structure of the electronic database requires utmost precision; therefore, any non-systemic items are not justified. Due to the above reasons, a lexeme, or a lexicographic unit in one of its meanings, has been chosen as the key structural unit of the Database of Contemporary Lithuanian Lexicon (DCLL, Dabartinės lietuvių leksikos duomenų bazė). To produce a most precise card of a lexeme is much more rational and simpler than that of the whole lexicographic unit. In the process, a number of analogous cases and other cases of non-systemic description of data should be taken into account. Moreover, it is hardly possible to consider them all. Thus an attempt will be made to finally produce the DCLL as a card index of lexemes rather than lexicographic units (their entries). Dictionary entries are to be produced by joining the lexemes automatically.

This paper shows how to reduce evaluation time for context inference. Probabilistic Context Inference has proven to be a good representation of the physical reality with uncertain or missing information, giving with the probability also a measure of the quality of information. As the inference complexity is very high, the complexity of the to be evaluated rule (representing a share of the real world) should be reduced as far as possible. Therefore we present an approach to select only relevant values of context types and to adapt this selection during its usage time. A short proof of concept indicates that both targets, reducing inference time and maintaining quality of information, can be reached with the proposed approach.

The Unified Modeling Language (UML) is the standard to specify the structure and behaviour of software systems. The created models are a constitutive part of the software specification that serves as guideline for the implementation and the test of software systems. In order to verify the functionality which is defined within the specification documents, the domain experts need to perform an acceptance test. Hence, they have to generate test cases for the acceptance test. Since domain experts usually have a low level of software engineering knowledge, the test case generation process is challenging and error-prone. In this paper we propose an approach to generate high-level acceptance test plans automatically from business processes. These processes are modeled as UML Activity Diagrams (ACD). Our method enables the application of an all-path coverage criterion to business processes for testing software systems.

Many component based systems and frameworks require the integration of external codes, for example, providing numerical functionalities. These numerical codes can be either sequential or parallelized, written in languages such as C, Fortran, Python, or Java. Frameworks provide support for workflow management, data management, using distributed computing resources, or a graphical user interface. Today, modern systems are based on Eclipse and OSGi or similar technologies. For many frameworks, tight integration of pre-existing or third-party code requires manual source code changes to add the specific component interfaces to such code. As this is error-prone and time consuming, especially when large code bases must be integrated, tool support for these steps becomes useful, or even necessary. Tool support for automatic integration of existing code (in different languages) comprises several sub-problems such as code analysis, code transformation, generation of wrapper code, generation of proper user interfaces, and others. In this paper, we focus on the aspect of modularization of existing Java OSGi workflow systems and present a new Eclipse-based tool which provides end-user support for the migration of previously unmodularized software into modules or components.

The paper introduces the results of an investigation into spoken Lithuanian based on the Corpus of Spoken Lithuanian and carried out in Vytautas Magnus University. In the course of investigation the interrogative utterances used in spontaneous language were classified according to function and structure. The quantitative analysis of the data accessed from the Corpus of Spoken Lithuanian was based on a special computer program CHILDES. The results have shown that spoken Lithuanian gives preference to closed interrogative utterances rather than open interrogative utterances. Closed interrogative utterances in most cases use no interrogative particles; utterances with interrogative particles are much less frequent. In the latter subgroup interrogative particles mostly occur in the utterance-initial (e.g. ar, gal etc.) rather than utterance-final position (e.g. taip, ne, ane etc.). In the class of open interrogative utterances adverbial modifier utterances prevail; in the latter subclass interrogative utterances of the adverbial modifier of place dominate.

Social network analysis investigates the structure of relations amongst social actors. A general approach to detect patterns of interaction and to filter out irregularities is to classify actors into groups and to analyze the relational structure between and within the various classes. The first part of this paper presents methods to define and compute structural network positions, i. e., classes of actors dependent on the network structure. In the second part we present techniques to visualize a network together with a given assignment of actors into groups, where specific emphasis is given to the simultaneous visualization of micro and macro structure.

This paper compares the output of two translation tasks. In an attempt to find out the extent to which students of translation can translate Arabic contextualized collocations into English properly, two conflicting views about carrying out a translation task are tested. The first holds that avoiding the use of a dictionary in test sessions, though not in translation classes, would save time and yield better translation products, whereas the second contends that recourse to a dictionary is unavoidable at any translation task, including tests. The two opposing views have their corollary in a similar dispute which has already been settled in favour of the mental lexicon rather than the dictionary (cf. Rangelova/Echeandia 2001). The results of this study defeat the first claim and run counter to Rangelova/Echeandia's findings, though obtained from qualitatively a different test setting.

In many areas of computer science ontologies become more and more important. The use of ontologies for domain modeling often brings up the issue of ontology integration. The task of merging several ontologies, covering specific subdomains, into one unified ontology has to be solved. Many approaches for ontology integration aim at automating the process of ontology alignment. However, a complete automation is not feasible, and user interaction is always required. Nevertheless, most ontology integration tools offer only very limited support for the interactive part of the integration process. In this paper, we present a novel approach for the interactive integration of ontologies. The result of the ontology integration is incrementally updated after each definition of a correspondence between ontology elements. The user is guided through the ontologies to be integrated. By restricting the possible user actions, the integrity of all defined correspondences is ensured by the tool we developed. We evaluated our tool by integrating different regulations concerning building design.

This paper describes an approach to detect unknown event patterns. In this context, an event is not only something that happens, but also something that can be analysed. This task is processed by a Complex Event Processing (CEP) engine. CEP is an emerging technology for detecting known patterns of events and correlating them to complex events in real-time. In order to reach the goal of finding unknown patterns, several known detection algorithms are discussed. In our work, we focus on discriminant analysis used for recognizing unknown patterns for the use case of credit card transactions and the fraud problem connected with this kind of payment. It is necessary to develop new methods of fraud detection and prevention because of the negative impacts for vendors and customers caused by credit card fraudsters at present. At the same time we would like to make provisions for the more sophisticated fraud methods that will occur in the future.

The language of written business communication is characterised by the extensive use of phraseology, not only in terms of collocations and idiomatic expressions, but also of standard phrases in prototypical business genres. In any case, the phraseological information should be included in business dictionaries (in the following referred to as BDs) in accordance with the planned dictionary functions. Hence, the selection and presentation of the phraseological information should be decided by the lexicographer on the basis of the user needs alone and not on the recommendations of the phraseological literature about lexicographical practice. In this paper, I will firstly explain why lexicography and phraseology, although closely associated in a large number of studies, are quite different disciplines, and how their shared interest for dictionary practice in general is based on radically different views. I will then discuss the dictionary functions of BDs and focus on a number of concepts featuring extensive phraseological solutions to show and argue that dealing with phraseology in BDs should always keep focus on dictionary functions.

Speech translation research has made significant progress over the years with many high-visibility efforts showing that translation of spontaneously spoken speech from and to diverse languages is possible and applicable in a variety of domains. As language and domains continue to expand, practical concerns such as portability and reconfigurability of speech come into play: system maintenance becomes a key issue and data is never sufficient to cover the changing domains over varying languages. In this paper, we discuss strategies to overcome the limits of today's speech translation systems. In the first part, we describe our layered system architecture that allows for easy component integration, resource sharing across components, comparison of alternative approaches, and the migration toward hybrid desktop/PDA or stand-alone PDA systems. In the second part, we show how flexibility and reconfigurability is implemented by more radically relying on learning approaches and use our English–Thai two-way speech translation system as a concrete example.

This paper presents an approach for combining web-based learning (WBL) with problem-based learning (PBL), aiming at computer aided learning (CAL) support for introductory lectures to Computational Linguistics. Contrary to most current learning paradigms, we neither follow the "platform + content"-approach of current learning management systems used for distance learning, nor focus on the collaborative aspects of PBL. Instead we propose a text-centred concept for individual learning featuring problem-based interactive learning applications (TIP).

The increasing interest in multilingual applications like speech-to-speech translation systems is accompanied by the need for speech recognition front-ends in many languages that can also handle multiple input languages at the same time. In this paper we describe a universal speech recognition system that fulfills such needs. It is trained by sharing speech and text data across languages and thus reduces the number of parameters and overhead significantly at the cost of only slight accuracy loss. The final recognizer eases the burden of maintaining several monolingual engines, makes dedicated language identification obsolete and allows for code-switching within an utterance. To achieve these goals we developed new methods forconstructing multilingual acoustic models and multilingual n-gram language models.

This paper presents a real-time Java hardware and software system for use in embedded applications. The system includes a multithreaded Java microcontroller, an adapted Java Virtual Machine (JVH) and an extended application programming interface (API). The microcontroller core is designed with a multithreaded pipeline and hardware support for real-time scheduling algorithms. The Java Virtual Machine consists of trap routines for complex Java byte-codes, real-time garbage collection and a class loader for creating real-time data structures. On a higher level, the application programming interface provides easy real-time thread management and access to hardware components like serial interfaces and timer/counter units.

Association rules are amongst the most important patterns that can be discovered using data mining. Their automatic discovery is supported by most, if not all, data mining software tools. Moreover, many techniques have been devised to lter the most interesting (in many senses) rules from the complete set of discovered rules, such that the users are not swamped by results. However, association rules are actually hard to understand; even more so if one only looks at the most interesting rules. For example, rather strong correlations between attributes are not always obvious from the discovered rules. Similarly, a deeper explanation of related association rules may be missing from the rule set. In this paper we show how Mosaic plots and, especially, their variant called Double Decker plots, can be used to visualize association rules. These plots visualize the contingency table that yields the association rule as well as the other potential rules in that table, whether they meet the thresholds or not. This gives a deeper understanding on the nature of the correlation between the left-hand side of the rule and the right-hand side. Moreover, we show how an interactive use of these plots helps the user to understand the relationship between related association rules.

This paper presents a context-dependent hybrid connectionist speech recognition system that uses a set of generalized hierarchical mixtures of experts (HME) to estimate context-dependent posterior acoustic class probabilities. The connectionist part of the system is organized in a modular fashion, allowing the distributed training of such a system on regular workstations. Context classes are based on polyphonic contexts, clustered using decision trees which we adopt from our continuous density HMM recognizer JANUS (Waibel et al., 1996). The system is evaluated on ESST, an English speaker-independent spontaneous speech database. Context dependent modeling is shown to yield significant improvements over simple context-independent modeling, requiring only small additional overhead in terms of training and decoding time.

Decoding algorithm is a crucial part in statistical machine translation. We describe a stack decoding algorithm in this paper. We present the hypothesis scoring method and the heuristics used in our algorithm. We report several techniques deployed to improve the performance of the decoder. We also introduce a simplified model to moderate the sparse data problem and to speed up the decoding process. We evaluate and compare these techniques/models in our statistical machine translation system.

This paper describes JANUS-III, our most recent version of the JANUS speech-to-speech translation system. We present an overview of the system and focus on how system design facilitates speech translation between multiple languages, and allows for easy adaptation to new source and target languages. We also describe our methodology for evaluation of end-to-end system performance with a variety of source and target languages. For system development and evaluation, we have experimented with both push-to-talk as well as cross-talk recording conditions. To date, our system has achieved performance levels of over 80% acceptable translations on transcribed input, and over 70% acceptable translations on speech input recognized with a 75-90% word accuracy. Our current major research is concentrated on enhancing the capabilities of the system to deal with input in broad and general domains.

This paper deals with buffer management for tree-based access-path structures in real-time database system kernels. A predictable and efficient buffer replacement strategy for B- and B/sup +/-trees is introduced. The strategy called LC (Level Control) uses the structural information known about the trees and the tree search to determine a node to be replaced. The basic idea is to use the level of a node as replacement criteria. This allows a certain prediction of buffer hits and misses for tree search operations, even if the item's value searched for is unknown. A small prediction interval can be calculated for search operations and a larger interval for update operations. Simulations show the correctness of these calculations and the efficiency of LC for B- and B/sup +/-trees compared to common other replacement strategies.

Building robust stochastic language models is a major issue in speech recognition systems. Conventional word-based n-gram models do not capture any linguistic constraints inherent in speech. In this paper, the notion of function and content words (open/closed word classes) is used to provide linguistic knowledge that can be incorporated into language models. Function words are articles, prepositions and personal pronouns. Content words are nouns, verbs, adjectives and adverbs. Based on this class definition resulting in function and content word markers, a new language model is defined. A combination of the word-based model with this new model is introduced. The combined model shows modest improvements both in perplexity results and recognition performance.

Inexpensive mobile data communication - based, e.g., on cellular phone networks - enables the access to data and services for "everyone, anytime, anywhere". At the same time, global information systems begin to emerge, which provide a unified access to multimedia data and services. However, the mobile access to these information repositories is being hampered by the technical constraints of the underlying hardware, especially with respect to interactive multimedia applications. Central challenges are low bandwidth in comparison to local-area networks, resource constraints of mobile devices, and massive changes in the information stock. Based on an integrative view of systems and data models for mobile information visualization, this paper discusses solution strategies which address these challenges. Furthermore, we discuss prototype applications being developed at our institute ("HyperFunk", "Active Multimedia Mail") and the basic research project "MOVI".

The simultaneous interpreter is a secondary text producer in the sense that s/he is not responsible for the contents of the text to be transferred to the target language. In this paper it is assumed 1) that this lack of autonomy results in the construction of a discourse model which differs from the speaker’s discourse model and 2) that this difference is reflected in different referential expressions used by the speaker and the interpreter respectively. Based on Mira Ariel’s Accessibility Theory, referential expressions are regarded as accessibility markers, by means of which the text producer signals to the hearer whether the mental representation of the referent in question has a high or low degree of accessibility in his/her memory. It is claimed that, whenever text production is difficult due to special conditions of the simultaneous mode, the interpreter tends to choose referential expressions which indicate that a particular referent has a lower state of activation in his/her discourse model as compared with the speaker’s. If a higher accessibility marker is used in such a case this can be explained as a misunderstanding, comprehension deficit or loss of information.