Utilizing social media in software documentation and knowledge management

This Master’s thesis studies the possibilities that social media tools can bring to help knowledge management in software development companies. It will introduce the most popular tools of social media and their usage possibilities in companies, not forgetting the possible downsides. One relevant aspect in this study is to investigate the possibilities of social media to help converting existing tacit knowledge into explicit. The purpose of the work is to create a proposal of social media utilization for a mid-sized software company, which has not utilized social media tools before. To be able to create the proposal, employees of the company are interviewed and a survey is executed to analyze the current situation. In addition a pilot project for trying out new social media tools is executed. The final result of this thesis introduces a tailored solution for the target company to start utilizing social media in its documentation and knowledge sharing processes. This new solution consists of multiple individual suggestions that are categorized and prioritized based on the significance and benefit that they bring to the company.

Kernel-Based Ranking. Methods for Learning and Performance Estimation

Machine learning provides tools for automated construction of predictive models in data intensive areas of engineering and science. The family of regularized kernel methods have in the recent years become one of the mainstream approaches to machine learning, due to a number of advantages the methods share. The approach provides theoretically well-founded solutions to the problems of under- and overfitting, allows learning from structured data, and has been empirically demonstrated to yield high predictive performance on a wide range of application domains. Historically, the problems of classification and regression have gained the majority of attention in the field. In this thesis we focus on another type of learning problem, that of learning to rank. In learning to rank, the aim is from a set of past observations to learn a ranking function that can order new objects according to how well they match some underlying criterion of goodness. As an important special case of the setting, we can recover the bipartite ranking problem, corresponding to maximizing the area under the ROC curve (AUC) in binary classification. Ranking applications appear in a large variety of settings, examples encountered in this thesis include document retrieval in web search, recommender systems, information extraction and automated parsing of natural language. We consider the pairwise approach to learning to rank, where ranking models are learned by minimizing the expected probability of ranking any two randomly drawn test examples incorrectly. The development of computationally efficient kernel methods, based on this approach, has in the past proven to be challenging. Moreover, it is not clear what techniques for estimating the predictive performance of learned models are the most reliable in the ranking setting, and how the techniques can be implemented efficiently. The contributions of this thesis are as follows. First, we develop RankRLS, a computationally efficient kernel method for learning to rank, that is based on minimizing a regularized pairwise least-squares loss. In addition to training methods, we introduce a variety of algorithms for tasks such as model selection, multi-output learning, and cross-validation, based on computational shortcuts from matrix algebra. Second, we improve the fastest known training method for the linear version of the RankSVM algorithm, which is one of the most well established methods for learning to rank. Third, we study the combination of the empirical kernel map and reduced set approximation, which allows the large-scale training of kernel machines using linear solvers, and propose computationally efficient solutions to cross-validation when using the approach. Next, we explore the problem of reliable cross-validation when using AUC as a performance criterion, through an extensive simulation study. We demonstrate that the proposed leave-pair-out cross-validation approach leads to more reliable performance estimation than commonly used alternative approaches. Finally, we present a case study on applying machine learning to information extraction from biomedical literature, which combines several of the approaches considered in the thesis. The thesis is divided into two parts. Part I provides the background for the research work and summarizes the most central results, Part II consists of the five original research articles that are the main contribution of this thesis.

Service innovation development and knowledge integration practices - a comparative case study on KIBS sector

Academic research on services and innovations on services has significantly grown during recent years. So far research concerning management of knowledge intensive work on service development activities is very limited. The objective of this study was to examine knowledge integration practices that support service innovation development and to the best of knowledge such studies have not been previously published in academic literature. In the theoretical part of the study a review of state‐of‐the‐art literature was conducted, research gap was indicated and a framework for analysis was built. In the empirical part an explorative comparative multi‐case study was carried out in KIBS sector. Four companies were selected and four service development projects were inspected. The service development activities and knowledge integration practices were identified. The cases were carefully compared and results formed. The empirical results indicated that service innovation development is partly linear and partly incremental flow of activities where knowledge integration practices have important role supporting the planning and execution of tasks. Knowledge integration practices supporting planning and workshops are close interaction, interpretation, project planning and sequencing of work tasks. The identified knowledge integration practices supporting building service solution were careful role and competence management, routines and common knowledge. The main implication is that to manage knowledge intensive service innovation development a firm should carefully develop and choose relevant knowledge integration practices to support the service development activities.

Lectio praecursoria : Systems between information and knowledge : in a memory management model of an extended enterprise

VTT Jouni Meriluodon valtio-opin alaan kuuluva väitöskirja Systems between information and knowledge : in a memory management model of an extended enterprise tarkastettiin 21.6.2011 Helsingin yliopistossa.

Local Embeddedness and Knowledge Acquisition and How They Affect Finnish Business Operations in Russia

This pilot project aims examine the factors of the Finnish subsidiaries local embeddedness, their knowledge creation capabilities and the transfer mechanisms of new practices in the context of the Russian market. The research is designed as a multiple case study conducted with a qualitative approach. The empirical data consists of the interviews of the four Finnish case companies operating in the Kaluga region and three local partner companies. The deductive and inductive approaches were employed to conduct the analysis of the data. The propositions for the future study were developed in the conclusive chapters of the research, where we propose that the factor of the economy growth and industrialization matters in terms of subsidiaries’ role dedication, their knowledge creation capabilities, and direction of the knowledge flow within the local environment.

The brain decade in debate: I. Neurobiology of learning and memory

This article is a transcription of an electronic symposium in which some active researchers were invited by the Brazilian Society for Neuroscience and Behavior (SBNeC) to discuss the last decade's advances in neurobiology of learning and memory. The way different parts of the brain are recruited during the storage of different kinds of memory (e.g., short-term vs long-term memory, declarative vs procedural memory) and even the property of these divisions were discussed. It was pointed out that the brain does not really store memories, but stores traces of information that are later used to create memories, not always expressing a completely veridical picture of the past experienced reality. To perform this process different parts of the brain act as important nodes of the neural network that encode, store and retrieve the information that will be used to create memories. Some of the brain regions are recognizably active during the activation of short-term working memory (e.g., prefrontal cortex), or the storage of information retrieved as long-term explicit memories (e.g., hippocampus and related cortical areas) or the modulation of the storage of memories related to emotional events (e.g., amygdala). This does not mean that there is a separate neural structure completely supporting the storage of each kind of memory but means that these memories critically depend on the functioning of these neural structures. The current view is that there is no sense in talking about hippocampus-based or amygdala-based memory since this implies that there is a one-to-one correspondence. The present question to be solved is how systems interact in memory. The pertinence of attributing a critical role to cellular processes like synaptic tagging and protein kinase A activation to explain the memory storage processes at the cellular level was also discussed.

Effects of caffeine on learning and memory in rats tested in the Morris water maze

We studied some of the characteristics of the improving effect of the non-specific adenosine receptor antagonist, caffeine, using an animal model of learning and memory. Groups of 12 adult male Wistar rats receiving caffeine (0.3-30 mg/kg, ip, in 0.1 ml/100 g body weight) administered 30 min before training, immediately after training, or 30 min before the test session were tested in the spatial version of the Morris water maze task. Post-training administration of caffeine improved memory retention at the doses of 0.3-10 mg/kg (the rats swam up to 600 cm less to find the platform in the test session, P<=0.05) but not at the dose of 30 mg/kg. Pre-test caffeine administration also caused a small increase in memory retrieval (the escape path of the rats was up to 500 cm shorter, P<=0.05). In contrast, pre-training caffeine administration did not alter the performance of the animals either in the training or in the test session. These data provide evidence that caffeine improves memory retention but not memory acquisition, explaining some discrepancies among reports in the literature.

Liber fostering Open Science and Knowledge Discovery

Presentation of Kristiina Hormia-Poutanen at the 25th Anniversary Conference of The National Repository Library of Finland, Kuopio 22th of May 2015.

Effects of isoflavone on the learning and memory of women in menopause: a double-blind placebo-controlled study

Hormone decline is common to all women during aging and, associated with other factors, leads to cognitive impairment. Its replacement enhances cognitive performance, but not all women present a clinical and family or personal history that justifies its use, mainly women with a history of cancer. The aim of this study was to determine whether a daily oral dose of 80 mg of isoflavone extract for 4 months can produce benefits in women with low hormone levels, contributing to improvement in cognitive aspects. The sample comprised 50- to 65-year-old women whose menstruation had ceased at least 1 year before and who had not undergone hormone replacement. The volunteers were allocated to two groups of 19 individuals each, i.e., isoflavone and placebo. There was a weak correlation between menopause duration and low performance in the capacity to manipulate information (central executive). We observed an increase in the capacity to integrate information in the group treated with isoflavone, but no improvement in the capacity to form new memories. We did not observe differences between groups in terms of signs and symptoms suggestive of depression according to the Geriatric Depression Scale. Our results point to a possible beneficial effect of isoflavone on some abilities of the central executive. These effects could also contribute to minimizing the impact of memory impairment. Further research based on controlled clinical trials is necessary to reach consistent conclusions.

Possible roles of COX-1 in learning and memory impairment induced by traumatic brain injury in mice

People who suffer from traumatic brain injury (TBI) often experience cognitive deficits in spatial reference and working memory. The possible roles of cyclooxygenase-1 (COX-1) in learning and memory impairment in mice with TBI are far from well known. Adult mice subjected to TBI were treated with the COX-1 selective inhibitor SC560. Performance in the open field and on the beam walk was then used to assess motor and behavioral function 1, 3, 7, 14, and 21 days following injury. Acquisition of spatial learning and memory retention was assessed using the Morris water maze on day 15 post-TBI. The expressions of COX-1, prostaglandin E2 (PGE2), interleukin (IL)-6, brain-derived neurotrophic factor (BDNF), platelet-derived growth factor BB (PDGF-BB), synapsin-I, and synaptophysin were detected in TBI mice. Administration of SC560 improved performance of beam walk tasks as well as spatial learning and memory after TBI. SC560 also reduced expressions of inflammatory markers IL-6 and PGE2, and reversed the expressions of COX-1, BDNF, PDGF-BB, synapsin-I, and synaptophysin in TBI mice. The present findings demonstrated that COX-1 might play an important role in cognitive deficits after TBI and that selective COX-1 inhibition should be further investigated as a potential therapeutic approach for TBI.