n this dissertation a comparison of two evolutionary methods for evolving ANNs for robot control is made. The methods compared are SANE with enforced sub-population and delta-coding, and marker-based encoding. In an attempt to speed up evolution, marker-based encoding is extended with delta-coding. The task selected for comparison is the hunter-prey task. This task requires the robot controller to posess some form of memory as the prey can move out of sensor range.

Vid analys av fluider med akustisk spektroskopi finns ett behov av att finna multivariata metoder för att utifrån akustiska spektra prediktera storheter såsom viskositet och densitet. Användning av artificiella neurala nätverk och bayesiska nätverk för detta syfte utreds genom teoretiska och praktiska undersökningar. Förbehandling och uppdelning av data samt en handfull linjära och olinjära multivariata analysmetoder beskrivs och implementeras. Prediktionsfelen för de olika metoderna jämförs och PLS (Partial Least Squares) framstår som den starkaste kandidaten för att prediktera de sökta storheterna..

Managing credit risk might be the single most important business area for any commercial bank. The assessment of "good" and "bad" corporate clients is a important task for a creditor. A bad debtor is a corporate client with hardships in meeting the continous claims (interest payments) that a creditor requires. One way of evaluating or separating a "bad" client from a "good" client is to assess the propensity for the client to file for bankruptcy. This thesis examines 226 firms in the Swedsh market in the quest of predicting corporate bankruptcy.

Managing credit risk might be the single most important business area for any commercial bank. The assessment of "good" and "bad" corporate clients is a important task for a creditor. A bad debtor is a corporate client with hardships in meeting the continous claims (interest payments) that a creditor requires. One way of evaluating or separating a "bad" client from a "good" client is to assess the propensity for the client to file for bankruptcy. This thesis examines 226 firms in the Swedsh market in the quest of predicting corporate bankruptcy.

Den klassiska approachen till navigering innefattar att agenten håller en intern representativ modell av omgivningen. Denna approach har emellertid många nackdelar, speciellt för dynamiska miljöer. En modernare approach är att förlita sig på den faktiska omgivningen istället för en modell av denna. Detta arbete presenterar en undersökning av navigeringsproblemet och hur väl det löses av agenter vars kontrollmekanismer utgörs artificiella neurala nätverk. Tillförlitligheten hos de två neurala arkitekturerna Extended sequential cascaded network och Self-organized recurrent network bestäms genom experiment.

AI är något som blir allt viktigare inom dagens spel och får allt högre krav på att agera mänskligt och intelligent. Detta kandidatarbete undersöker vilken metod som är att föredra för att skapa en AI som kan lära sig av sina tidigare erfarenheter. Några av de metoder som undersöks är trädstrukturer, Artificial Neural Network och GoCap. Genom att skapa en applikation med en av metoderna samt göra en undersökning på hur AI:n i applikationen upplevdes fick vi resultat om denna metod var användbar. Utifrån detta diskuteras det ifall andra metoder hade varit mer effektiva, hur man hade kunnat förbättra AI:n samt hur framtiden för spel-AI skulle kunna se ut..

Denna rapport jämför prestandan mellan state-of-the-art AI-botar i racing spelet TORCS och en AI-bot som kör med hjälp av ett artificiellt neuralt nätverk (ANN-bot). ANN-boten, som implementerades som en del av arbetet, använder en feedforward arkitektur och backpropagation för inlärning. Ett separat program som användes för att träna det neurala nätverket med träningdata som spelats in från TORCS implementerades också. Som state-of-the-art AI-botar användes AI-botar som har använts i en tävling. De fyra AI-botarna testades på åtta olika banor och data om hur lång tid varje varv tog och hur snabbt AI-botarna körde sparades och sammanställdes. Resultaten visar att på banorna som ANN-boten klarar av att köra runt så är ANN-boten snabbare än en den långsamaste state-of-the-art boten, men ANNboten klara inte av majoriteten av banorna som den testades på.

This thesis presents a system which has been implemented to satisfy a need in theresearch on how speech planning interacts with syntactic and prosodic structure inspontaneous speech. The long-term purpose of the research is to provide models forautomatic parsing of spontaneous speech and for psycholinguistical modelling of speechproduction. Identification of inhalation pauses is an important step in the developmentof automatic methods for spontaneous speech parsing.Identification of inhalation pauses is considered to be a keyword-spotting speechrecognition problem. Hybrid HMM(Hidden Markov Models)/ANN(Artificial NeuralNetworks) approach is applied to this problem. Method gets 90,8% in Recall, 66,4% inPrecision and 76,7% in F-score.

The purpose for this Masters thesis is to describe childrens librarians work in supplying literature to children with emphasis on easy-reader materials. Our main questions are: · What do librarians believe characterizes an easy-reader book, and what do they base their decision on when classifying a book as easy-to-read? · How have the libraries in our study solved the hands-on difficulties associated with easy readers childrens literature? · Which factors do librarians consider to be important when supplying literature to children that are not used to reading or are not good readers? The theories we have used are Aidan Chambers Reading Circle and Karner Smidts view on the librarians professional roles when supplying literature. We have used a case study method, which means that we have gathered information through observation, objective papers and interviews. This has been done at six libraries.

AI är något som blir allt viktigare inom dagens spel och får allt högre krav på att agera mänskligt och intelligent. Detta kandidatarbete undersöker vilken metod som är att föredra för att skapa en AI som kan lära sig av sina tidigare erfarenheter. Några av de metoder som undersöks är trädstrukturer, Artificial Neural Network och GoCap. Genom att skapa en applikation med en av metoderna samt göra en undersökning på hur AI:n i applikationen upplevdes fick vi resultat om denna metod var användbar. Utifrån detta diskuteras det ifall andra metoder hade varit mer effektiva, hur man hade kunnat förbättra AI:n samt hur framtiden för spel-AI skulle kunna se ut..

According to the study [1] several firefighters are killed in action each year. This happens naturally in various ways. One of the major factors is that firefighters often do not know how the environment looks like, for example that a building in an industrial area contains hazardous gas cylinders. Without a solution, fire-fighters will continue to die in duty. It is therefore important to solve this problem so that we can minimize the risks and save lives.The goal of this project was a control system for pneumatic artificial muscles.

Since the brain is an important prerequisite for human language there is a great interest to gain more knowledge about healthy brain activity during language mediated communication. This study examines variations in high level language ability relating to demographic factors like gender and age and relates language ability to neural activity.Eighteen individuals in ages between 22 and 64 were included and divided into groups in relation to gender and age. The material used to assess language ability was derived from Testbatteri för Bedömning av Subtila Språkstörningar (Laakso, Brunnegård, Hartelius & Ahlsén, 2000), Swedish Lexical Decision Test (Almkvist, Adveen, Henning & Tallberg, 2007) and a reading test (Högskoleprovet). The results were correlated with measured brain activity using functional Magnetic Resonance Imaging (fMRI) and language paradigms. ANOVA was applied to discover possible demographic variances in language ability as well as in brain activity.The results showed no significant differences in language ability.

The tourism industry is growing and with its growth competition between destinations rises. In competition, cooperation and collaboration between tourism operators seem beneficial when creating an attractive and dynamic tourist destination.The purpose of this study is to describe and discuss the significance and influence of tourist related networks at a tourist destination, and therefore this study aims to examine the question: What significance do the tourist related networks have at a tourist destination?This study focuses on seven networks on Gotland that are tourism related, and the representatives of the networks were interviewed in order to get their view on the network and its significance to the destination. The study shows that networks have an important role in creating a unified tourist product of the destination, and in reaching out to the potential visitors through joint destination marketing. The study also shows that the tourism related networks helps to improve the reception of the tourist at a destination and to secure the quality of the tourist destination and its tourist products..

The aim of our study was to examine how Swedish journalists use the three social networks Facebook, Twitter and blogs to find news. We wanted to study which group of journalists that were using these networks most frequently and why. We also wanted to find out which of the three networks they use the most.We made a questionnaire which we sent out by e-mail to 400 Swedish journalists. 215 of them participated in our study.We analyzed the results of the responses from the questionnaire and discussed patterns and relations between them.The study establishes that the majority of the respondents believe that the networks are very useful and effective in their work.Our results show that out of the three social networks the journalists who participated in our study prefer to use blogs as sources in their line of work, as they are full of information from individuals and contain unique facts and stories.Alongside the results we discussed how social networks might be used in the future of journalism. Most of the respondents are aware of the possibility that they may need to increase the usage of them eventually, and that they will be an important part of the journalistic work process in the future. .

Denna rapport jämför prestandan mellan state-of-the-art AI-botar i racing spelet TORCS och en AI-bot som kör med hjälp av ett artificiellt neuralt nätverk (ANN-bot). ANN-boten, som implementerades som en del av arbetet, använder en feedforward arkitektur och backpropagation för inlärning. Ett separat program som användes för att träna det neurala nätverket med träningdata som spelats in från TORCS implementerades också. Som state-of-the-art AI-botar användes AI-botar som har använts i en tävling. De fyra AI-botarna testades på åtta olika banor och data om hur lång tid varje varv tog och hur snabbt AI-botarna körde sparades och sammanställdes.