Algirdas Laukaitis

Social systems in terms of coherent individual neurodynamics: conceptual premises, experimental and simulation scope

After reviewing numerous theories and experiments, our research adopted the field-theoretical deductive approach to shed new light on complex social systems as coherent neurodynamic processes taking place in individual minds. In this interdisciplinary study, we have outlined some general fundamental design principles of the field-theoretical view of the oscillating agent as well as of coherent social systems. From the systems point of view, ordered social systems by their own intrinsic nature are interpreted as coherent activations via the mind-field medium of social agents. Consequently, this study not only provides the major conceptual assumptions of the proposed (Oscillation-Based Multi-Agent System [OSIMAS]) paradigm but also presents an electroencephalography-based inductive experimental validation framework and some empirical results to validate major OSIMAS assumptions. Based on the conceptual and experimental findings, we constructed modelling framework and presented oscillations-based micro (coupled oscillator energy exchange model) and macro (MEPSM1) simulation models. We also systemized some other studies and applications, which are most relevant to the work presented here.

Formal concept analysis and information systems modeling

In this paper, we present the methodology based on formal concept analysis (FCA) for the information system model verification. We show that FCA can be useful for understanding conceptual model topology and it can be used to improve conceptual model hierarchal structure. IBMs Information Framework (IFW) Financial Services Data Model (FSDM) has been used for the present research. By using FSDM we demonstrate that the IS model can be recreated with formal concept analysis.

Natural language processing and the conceptual model self-organizing map

Self-organizing map can be an effective tool for the textual data classification. In this paper, we represent the methodology of an integration of the information system modeling and the development of the information system natural language interface. The main idea of the paper is to build the set of self-organising maps from information system documentation and then reuse it in human-machine communication as a semantic parsing component. The IBMs Information Framework (IFW) Financial Services Data Model has been used in an experiment where we tested how appropriate is presented methodology and what is classification accuracy of the received self-organizing maps. We compare classification accuracy with the IBMs WebSphere Voice Server NLU solution and demonstrate that self-organising maps can be a competitive components in the information systems natural language interfaces.

Jmining — Information Delivery WEB Portal Architecture and Open Source Implementation

In this paper we present a new framework and its Java based open source implementation for building web information delivery portals (IDP). The framework provides the necessary infrastructure for development and management of web portal using only web browser interface. Presented framework enables speedup of information delivery portal development compared with some commercial available solutions. All presented components are implemented as separate open source project JMining (VGTU Java open source project). The contribution of this paper is threefold: Firstly, we introduce a IDP framework based on atomic application objects container. Proposed framework is characterized by its simplicity and a possibility to build complex information delivery web portals. Secondly, we show new way of deploying applications into portal using only web browsers. We discuss new requirements for web browsers and show that up today only few of them meets them. Thirdly, all presented concepts are implemented as Java open source project. We present discussion about open source projects and importance for support for such projects from academic environment. Our research shows that until now there was no open source project in information delivery portals domain and we think that our project can fulfil such gap.

Towards oscillations-based simulation of social systems: a neurodynamic approach

This multidisciplinary work presents synopsis of theories in the search for common field-like fundamental principles of self-organisation and communication existing on quantum, cellular, and even social levels. Based on these fundamental principles, we formulate conceptually novel social neuroscience paradigm (OSIMAS), which envisages social systems emerging from the coherent neurodynamical processes taking place in the individual mind-fields. In this way, societies are understood as global processes emerging from the superposition of the conscious and subconscious mind-fields of individual members of society. For the experimental validation of the biologically inspired OSIMAS paradigm, we have designed a framework of EEG-based experiments. Initial baseline individual tests of spectral cross-correlations of EEG-recorded brainwave patterns for some mental states have been provided in this paper. Preliminary experimental results do not refute the main OSIMAS postulates. This paper also provides some insights for the construction of OSIMAS-based simulation models.

Automation of Merging in ERP Revision Control

In this paper, we describe a model for extracting rules that describe enterprise resource planning (ERP) system upgrade process. The rules are extracted automatically by analyzing programming code from completed ERP systems upgrade projects. Later those rules are verified and tuned by experienced programmer. The rules that we are extracting are described in the language equivalent to the first-order logic (i.e. expressivity of Turing machine). Nevertheless, we put a constrain on the rules description language by defining how knowledge base is used to define these rules. We require that ERP system code and ERP upgrade knowledge base must be transformed to a series of aligned strings without lost of expressivity. Such strong requirement ensures that we are able to use existing machine learning algorithms in the process of software development and upgrade. These series of strings are compared by strings manipulation algorithms and then differences are resolved by merge algorithm presented in this paper. We used Microsoft Dynamics NAV as an example to test usefulness of presented method but other systems can be used as well if they can be presented as series of code strings.

Application of the program agent, that uses realty ontology, as the data source for the decision support system

Santrauka Siekiant didesnio darbo nasumo žinių visuomenėje, greitas ir kokybiskas informacijos gavimas tampa svarbiu faktoriumi, kuris lemia organizacijos sėkme atkaklioje konkurencinėje kovoje. Norint pasiekti užsibrėžte tikslė, siuo metu jau nebeužtenka naudotis informacijos paieskos internete sistemomis, tokiomis kaip Google ar Yahoo. Siame straipsnyje pateikiama programinio agento architektūra ir jo realizacijos metu atlikti tyrimai, kurie parodo, kad siėlomi sprendimai leidžia surinkti ir sutvarkyti nekilnojamojo turto skelbimus daug efektyviau, nei tai darytų žmogus, naudojantis tik interneto informacijos paieskos sistemas. Nekilnojamojo turto rinka pasirinkta neatsitiktinai. Tai viena is dinamiskiausiė rinkų Lietuvoje, kurioje greitas informacijos pateikimas lemia sėkmingus sprendimus. Nors pateikti sprendimai ir orientuoti į nekilnojamojo turto segmentė, taėiau daugelis koncepcijų gali būti sėkmingai pritaikyti ir kitiems ekonomikos segmentams. Pagrindinis akcentas pateiktoje programinio agento ar...

Asymmetric Hybrid Machine Translation for Languages with Scarce Resources

Half of the world speaks languages that are out of the machine translation and natural language processing technologies mainstream. Then the choice of natural language technology for a given language pair is greatly impacted by technology and resources available. In this paper we describe a hybrid architecture and technology for rapid development of the machine translation system from English to low-density languages. We use state of the art English language processing technologies and resources to transform (compress) the language into the more abstract form. The abstraction level of the transformation is adapted to our knowledge of the low-density (foreign) language. Then statistical machine translation is used to induce translation rules. All tests and implementations have been done on the English --- Lithuanian language pair. Some of the findings of the research can be useful for all, novel and old machine translation language pairs.

Automation of Upgrade Process for Enterprise Resource Planning Systems

This paper presents a framework for semi-automatic process of enterprise resource planning (ERP) system upgrade. We suggest to change currently accepted practice of manual upgrade process when domain expert-programmer works through all localizations and transforms them manually to the new version of ERP system. The core idea for this framework is to induce the software code transformation patterns from completed upgrade projects and then to refine these patterns by using knowledge of ERP upgrade expert. These patterns lets us to increase productivity of upgrade process by improving automatic code alignment and annotation and by providing code transformation to the new version of ERP system. The price for these improvements is a requirement for upgrade expert to move from traditional 4/GL ERP programming language to stochastic meta-programming language which is used to describe code alignment and code transformation patterns.

Natural language based heavy personal assistant architecture for information retrieval and presentation

In this paper we present the progress of the natural language usage as the paradigm for information extraction and presentation in the enterprise environment. Distributed heavy personal assistant architecture and its implementation is presented as the solution to overcome difficulties related of the natural language use in the information systems development. A new methodology based on connectionist and symbol processing techniques for a knowledge worker to process his documents and utterance is suggested. Then we suggest the results from those processes to reuse for new documents classification and generation of small atomic applications. Finally the experiment is presented. We compare Microsoft EQ, IBM WebSphere Voice Server NLU toolbox and our solution for concepts identification accuracy.