Wojciech Korczynski

Agent-Oriented Computing Platform in Python

In this paper we present an architecture and implementation details of an agent-based computing and simulation platform. The introduced system is built based on easily exchangeable components that may be replaced (also in runtime). Such flexibility may be treated as a significant advantage in the case of decision support systems which often have to adapt themselves to varying constraints and trends and where a quick approximated options frequently have to be provided in a possibly shortest time.

Agent-oriented Foraminifera Habitat Simulation

An agent-oriented software solution for simulation of marine unicellular organisms called foraminifera is presented. Their simplified microhabitat interactions are described and implemented to run the model and verify its flexibility. This group of well fossilizable protists has been selected due to its excellent “in fossilio” record that should help to verify our future long-run evolutionary results. The introduced system is built utilizing PyAge platform and based on easily exchangeable components that may be replaced (also in runtime). Selected experiments considering substantial and technological efficiency were conducted and their results are presented and discussed. We show that our artificial life simulation follows realistic rules known from nature. The Lotka-Volterra predator-prey dynamics was simulated that proves the potential for extendability to more complex biological phenomena.

Enhancing Particle Swarm Optimization with Socio-cognitive Inspirations

We incorporate socio-cognitively inspired metaheuristics, which we have used successfully in the ACO algorithms in our past research, into the classical particle swarm optimization algorithms. The swarm is divided into species and the particles get inspired not only by the global and local optima, but share their knowledge of the optima with neighboring agents belonging to other species. Our experimental research gathered for common benchmark functions tackled in 100 dimensions show that the metaheuristics are effective and perform better than the classic PSO. We experimented with various proportions of different species in the swarm population to find the best mix of population.

Efficient Memetic Continuous Optimization in Agent-based Computing

This paper deals with a concept of memetic search in agent-based evolutionary computation. In the presented approach, local search is applied during mutation of an agent. Using memetic algorithms causes increased demand on the computing power as the number of fitness function calls increases, therefore careful planning of the fitness computing (through the proposed local search mechanism based on caching parts of the fitness function) leads to significant lowering of this demand. Moreover, applying local search with care, can lead to gradual improvement of the whole population. In the paper the results obtained for selected high-dimensional (5000 dimensions) benchmark functions are presented. Results obtained by the evolutionary and memetic multi-agent systems are compared with classic evolutionary algorithm.

Does Topic Modelling Reflect Semantic Prototypes

The chapter introduces a representation of a textual event as a mixture of semantic stereotypes and factual information. We also present a method to distinguish semantic prototypes that are specific for a given event from generic elements that might provide cause and result information. Moreover, this chapter discusses the results of experiments of unsupervised topic extraction performed on documents from a large-scale corpus with an additional temporal structure. These experiments were realized as a comparison of the nature of information provided by Latent Dirichlet Allocation based on Log-Entropy weights and Vector Space modelling. The impact of different corpus time windows on this information is discussed. Finally, we try to answer if the unsupervised topic modelling may reflect deeper semantic information, such as elements describing given event or its causes and results, and discern it from factual data.

Lamarckian and Lifelong Memetic Search in Agent-Based Computing.

Memetic algorithms when used with care can help in balancing exploitation and exploration of the metaheuristics, without the overhead measured by the rapidly increased number of function fitness calls. The paper tackles such balancing of use of metaheuristics in an agent-oriented setting. In particular, application of local search during a computing agent’s life is researched. The results shown for selected benchmark functions are presented along with necessary statistic testing.

ADAPTING A CONSTITUENCY PARSER TO USER-GENERATED CONTENT IN POLISH OPINION MINING

The paper focuses on the adjustment of NLP tools for Polish; e.g., morphological analyzers and parsers, to user-generated content (UGC). The authors discuss two rule-based techniques applied to improve their efficiency: pre-processing (text normalization) and parser adaptation (modified segmentation and parsing rules). A new solution to handle OOVs based on inflectional translation is also offered.

Retrieval and interpretation of textual geolocalized information based on semantic geolocalized relations

This paper describes a method for geolocalized information retrieval from natural language text and its interpretation by assigning them geographic coordinates. A proof-of-concept implementation is discussed, along with geolocalized dictionary stored in PostGIS / PostgreSQL spatial relational database. Discussed research focuses on strongly inflectional Polish language, hence additional complexity had to be taken into account. Presented method has been evaluated with the use of diverse metrics.

Extracting semantic prototypes and factual information from a large scale corpus using variable size window topic modelling

In this paper a model of textual events composed of a mixture of semantic stereotypes and factual information is proposed. A method is introduced that enables distinguishing automatically semantic prototypes of a general nature describing general categories of events from factual elements specific to a given event. Next, this paper presents the results of an experiment of unsupervised topic extraction performed on documents from a large-scale corpus with an additional temporal structure. This experiment was realized as a comparison of the nature of information provided by Latent Dirichlet Allocation and Vector Space modelling based on Log-Entropy weights. The impact of using different time windows of the corpus on the results of topic modelling is presented. Finally, a discussion is suggested on the issue if unsupervised topic modelling may reflect deeper semantic information, such as elements describing a given event or its causes and results, and discern it from pure factual data.