Universities worldwide face challenges both with students dropping out of educational programmes and repetitive questions directed toward teaching staff which both consume resources and result in delays. Recent progress in natural language processing (NLP) introduces the possibility of more sophisticated dialogue systems that could help alleviate the situation. Dialogue systems in education are complex to construct for multiple reasons. Two such reasons are that domain-specific data is often not readily available and extending an existing system often requires configuring the system again and re-training models. In this thesis, a graph-based knowledge base (KB) which is the foundation of a heavily rule-based dialogue system is proposed. The core of the natural language understanding (NLU) in the pipeline-based dialogue system includes the transformer-based DIET classifier for intent classification and entity extraction. The custom logic of the dialogue system relies on contextual and distributional embeddings. While the proposed solution is used in electrical engineering specifically, the KB and the architecture of the dialogue system are designed with generalization in mind. An emphasis is to maintain a low level of system maintenance after deployment allowing teaching staff without expertise in computer science and machine learning to operate the system. The utilization of transfer learning with pre-trained language models helps achieve this objective. The findings suggest that the system is sufficiently sophisticated to improve learning environments for students while potentially alleviating the workload of teaching staff. They further indicate that computer science and machine learning expertise are not required to operate the system over time. / Universitet världen runt möter utmaningar vad gäller både studenter som avbryter sina studier i förtid och repetitiva frågeställningar riktade till kursansvariga vilket konsumerar resurser och resulterar i onödig tidsåtgång. Den utveckling som på senare tid har skett inom naturlig språkhantering (NLP) introducerar möjligheter för mer sofistikerade dialogsystem som skulle kunna avhjälpa situationen. Dialogsystem inom utbildning är ofta komplexa att konstruera av flera anledningar. Två av dessa anledningar är att domän-specifik data sällan finns tillgänglig och att vidareutveckla existerande dialogsystem ofta kräver omkonfigurering och att man åter tränar de involverade modellerna. I denna uppsats föreslås en grafbaserad kunskapsbas (KB) som är grunden av ett till stora delar regelbaserat dialogsystem. Kärnan av den naturliga språkförståelsen (NLU) i det pipeline-baserade dialogsystemet inkluderar den transformer-baserade DIET-modellen för klassificering av intentioner och extrahering av entiteter. Den egenutvecklade logiken i dialogsystemet förlitar sig på förtränade kontextuella och distribuerade inbäddningar. Medan den föreslagna lösningen används specifikt inom elektroteknik så är både KB och dialogsystemets arkitektur utvecklade med generalisering i åtanke. Det finns även en betoning på att bibehålla en låg underhållningsnivå efter att systemet har sjösatts för att tillåta att systemet drivs av kursansvariga utan expertis inom datalogi eller maskininlärning. Användandet av förtränade språkmodeller hjälper till att uppnå detta mål. Upptäckterna tyder på att systemet är tillräckligt sofistikerat för att förbättra lärandemiljön för studenter medan det samtidigt möjligtvis kan hjälpa till att förminska arbetsbelastningen för kursansvariga. Vidare så indikerar upptäckterna att expertis inom datalogi och maskininlärning inte är nödvändigt för att driva systemet över tid.

The thesis investigates the view and knowledge among large versus small and medium sized enterprises (SME) regarding the implementation of the Corporate Sustainability Reporting Directive (CSRD). To enable a broad collection of data the study carries out semi-structured interviews and surveys on companies and stakeholders within the Swedish energy sector, as well as a literature review. The aim of the thesis is to identify the expected strengths, weaknesses, opportunities and threats and the knowledge base within the Swedish energy sector. Furthermore, the aim is to analyze potential consequences of implementation and to suggest potential mitigations where possible. The results indicate that large companies and SMEs to a great extent identify the same challenges and possibilities. However, large companies on average have a higher knowledge level than SMEs. The potential reasons behind this include differences in reporting time frame, level of engagement among sustainability responsible roles, lack of specific ESRS and difficulty in applying the directive. The greatest challenges relate to information dissemination in general and internally, information and data collection along the value chain, implementation for SMEs, increasing the knowledge level among companies and the structure of the directive. The potential solutions include different types of systemic structuring, collaborative efforts and extended responsibilities for stakeholders and authorities. Overall, the directive will entail positive changes and implement CSR as a statutory aspect, however, the directive still requires improvement to fully reach its’ desired potential. / I avhandlingen undersöks synen på och kunskapen om implementeringen av Corporate Sustainability Reporting Directive (CSRD) bland stora kontra små och medelstora företag (SME) i den svenska energisektorn. För att möjliggöra en bred insamling av data grundas studien på semistrukturerade intervjuer och enkäter hos företag, intervjuer med intressenter och en litteraturstudie. Syftet med avhandlingen är att identifiera förväntade styrkor, svagheter, möjligheter och hot samt kartlägga kunskapsbasen inom den svenska energisektorn. Vidare är syftet att analysera potentiella konsekvenser av implementeringen och föreslå potentiella åtgärder där det är möjligt. Resultaten visar att företag oavsett storlek i stor utsträckning identifierar samma utmaningar och möjligheter. Inom stora företagen finns dock i genomsnitt en högre kunskapsnivå än hos de små och medelstora företagen. De potentiella orsakerna till detta är bland annat skillnader i tidsramar för rapportering, graden av engagemang hos hållbarhetsansvariga, brist på specifika ESRS och svårigheter att tillämpa direktivet. De största utmaningarna gäller informationsspridning externt och internt, information och datainsamling inom värdekedjan, genomförande för små och medelstora företag, ökad kunskapsnivå bland företag och direktivets struktur. De potentiella lösningarna inkluderar olika typer av systemstrukturering, insatser och ansvarskrav för berörda parter och myndigheter. På det hela taget kommer direktivet att medföra positiva förändringar och göra företagens hållbarhetsansvar till en lagstadgad aspekt, men direktivet behöver fortfarande förbättras för att fullt ut nå sin önskade potential.

Description logics in their standard setting only allow for representing and reasoning with crisp knowledge without any degree of uncertainty. Of course, this is a serious shortcoming for use cases where it is impossible to perfectly determine the truth of a statement. For resolving this expressivity restriction, probabilistic variants of description logics have been introduced. Their model-theoretic semantics is built upon so-called probabilistic interpretations, that is, families of directed graphs the vertices and edges of which are labeled and for which there exists a probability measure on this graph family. Results of scientific experiments, e.g., in medicine, psychology, or biology, that are repeated several times can induce probabilistic interpretations in a natural way. In this document, we shall develop a suitable axiomatization technique for deducing terminological knowledge from the assertional data given in such probabilistic interpretations. More specifically, we consider a probabilistic variant of the description logic EL⊥, and provide a method for constructing a set of rules, so-called concept inclusions, from probabilistic interpretations in a sound and complete manner.

We consider ontology-based query answering in a setting where some of the data are numerical and of a probabilistic nature, such as data obtained from uncertain sensor readings. The uncertainty for such numerical values can be more precisely represented by continuous probability distributions than by discrete probabilities for numerical facts concerning exact values. For this reason, we extend existing approaches using discrete probability distributions over facts by continuous probability distributions over numerical values. We determine the exact (data and combined) complexity of query answering in extensions of the well-known description logics EL and ALC with numerical comparison operators in this probabilistic setting. / This is an extended version of the article in: Proceedings of the 11th International Symposium on Frontiers of Combining Systems. This version has been revised based on the comments of the reviewers.

For a probabilistic extension of the description logic EL⊥, we consider the task of automatic acquisition of terminological knowledge from a given probabilistic interpretation. Basically, such a probabilistic interpretation is a family of directed graphs the vertices and edges of which are labeled, and where a discrete probabilitymeasure on this graph family is present. The goal is to derive so-called concept inclusions which are expressible in the considered probabilistic description logic and which hold true in the given probabilistic interpretation. A procedure for an appropriate axiomatization of such graph families is proposed and its soundness and completeness is justified.

Formulae of linear temporal logic (LTL) can be used to specify (wanted or unwanted) properties of a dynamical system. In model checking, the system’s behaviour is described by a transition system, and one needs to check whether all possible traces of this transition system satisfy the formula. In runtime verification, one observes the actual system behaviour, which at any point in time yields a finite prefix of a trace. The task is then to check whether all continuations of this prefix to a trace satisfy (violate) the formula. More precisely, one wants to construct a monitor, i.e., a finite automaton that receives the finite prefix as input and then gives the right answer based on the state currently reached. In this paper, we extend the known approaches to LTL runtime verification in two directions. First, instead of propositional LTL we use the more expressive temporal logic ALC-LTL, which can use axioms of the Description Logic (DL) ALC instead of propositional variables to describe properties of single states of the system. Second, instead of assuming that the observed system behaviour provides us with complete information about the states of the system, we assume that states are described in an incomplete way by ALC-knowledge bases. We show that also in this setting monitors can effectively be constructed. The (double-exponential) size of the constructed monitors is in fact optimal, and not higher than in the propositional case. As an auxiliary result, we show how to construct Büchi automata for ALC-LTL-formulae, which yields alternative proofs for the known upper bounds of deciding satisfiability in ALC-LTL.

We propose an interactive repair method for the description logic EL that is based on the optimal-repair framework. The obtained repair might not be optimal in the theoretical sense, i.e. more than a minimal amount of consequences might have been removed---but from a practical perspective it is superior to a theoretically optimal repair as the interaction strategy enables the users to identify further faulty consequences connected to the initially reported errors. / This is an extended version of an article accepted at the KRR track of the 40th ACM/SIGAPP Symposium On Applied Computing (SAC 2025).

This study was motivated by concerns within the medical education community regarding the psychometric soundness of current assessment methodologies. More specifically, there is reason to seriously question the reliablity and/or validity of these methodologies in assessing the intellectual skills upon which medical competence is based.

This master's thesis describes Semantic Enrichment Component (SEC), that searches entities (e.g., persons or places) in the input text document and returns information about them. The goals of this component are to create a single interface for named entity recognition tools, to enable parallel document processing, to save memory while using the knowledge base, and to speed up access to its content. To achieve these goals, the output of the named entity recognition tools in the text was specified, the tool for storing the preprocessed knowledge base into the shared memory was implemented, and the client-server scheme was used to create the component.

企業透過創新以積極回應環境變遷的要求。企業新產品或服務的創新程度應依環境的需求決定。面對環境變遷，顧客需求可能改寫，企業如何取得正確的顧客需求成為重要問題；相反的，身處顧客需求已知的環境，競爭的重點轉為產品的品質與成本，企業內部如何充分分享知識以提升產品品質與降低成本成為重要議題。成功創新的企業是「如何」採行知識網路定位獲取所需要的新顧客需求及產品品質與成本等創新驅動因素，提升知識基礎與知識策略並進而影響創新能力？知識網路定位、知識基礎與知識策略對於創新能力的影響「有多大」？台灣資訊電子業又是如何透過知識網路定位同時達成上述的知識處理過程，提升創新能力？本研究試圖回答這些問題。 企業回應環境變遷所採行的企業策略可分為積極回應環境的前瞻者〈Prospector〉、固守環境穩定的防衛者〈Defender〉以及兼具前二者部份特色、試圖極大化獲利機會並同時極小化風險的分析者〈Analyzer〉(Miles et al., 1978)。企業回應環境的創新能力從本身具備的技術基礎與滿足市場需求所需具備技術間的差異程度可區分為激進式創新以及漸進式創新。文獻中對於企業創新能力的探討，主要有知識策略以及知識基礎二大觀點，知識策略強調企業回應環境需求變遷程度，設定新產品或服務創新程度、所需的激進式或漸進式創新能力，進而決定投資資源於探索新知識以及運用既有知識專案的比例，經由探索策略以及運用策略，提升企業回應環境需求的創新能力；知識基礎則強調企業的人力資本、組織資本以及社會資本等知識存量是提升企業創新能力的關鍵因素。如何調和知識基礎與知識策略產生正向的交互作用是提升企業創新能力的重要議題 (Cook & Brown, 1999)。 本研究檢視上述相關文獻，指出社會資本與人力資本、組織資本特性上的差異。鑲嵌不同社會資本〈新市場機會及信任〉的不同知識網路定位〈相對創業型定位及相對緊密型網路〉在環境變遷中提供企業不同程度的取得新市場機會〈如相對創業型定位有助於企業了解新市場顧客需求〉，促進知識基礎與知識策略對於企業創新能力的提升。本研究並且進一步在分析者企業策略的框架下，提出創新能力為因變數、知識策略與知識基礎為自變數、知識網路定位為調節變數的研究假設模型。 接著，本研究進行質性及量化實證研究，質性研究是從個案研究資料庫找出研究個案數量相對多且品質佳的創新個案：1980年代製造業的昇陽二號以及1990年代服務業的亞馬遜網路書店，逐篇檢視找出對應相關變數及其間關係，以初步檢驗本研究的研究模型。接著，為一般化本研究模型的解釋力，從重要的量化實證文獻中找出衡量各變數的問項整理成問卷，以台灣資訊電子業上市企業共341家為母體樣本，寄發問卷給各家企業之經理人，共取得69份有效問卷〈有效回收率為20.23%〉，將這些問卷資料以淨最小平方法演算後分析發現，採行分析〈者〉企業策略的台灣資訊電子業上市企業的實證資料支持絕大部分研究假設。在著重漸進式創新、漸進式創新專案對於激進式創新專案人力資本有顯著排擠效果的設計製造代工思維下，企業的知識網路定位透過知識策略與知識基礎提升創新能力。 本研究基於實證結果，除在理論上以知識網路定位銜接知識基礎與知識策略顯著提升對於創新能力的解釋力以及從產業創新階段觀點解答目前文獻對於組織知識網路定位的矛盾：同時採行相對創業型定位及相對緊密型網路成效差 (e.g., Koka & Prescott, 2008) vs. 同時採行二種知識網路定位成效好 (e.g., Capaldo, 2007) 外，同時也對實務提出建言：經理人應觀察產業所處的創新階段決定所需要提升的創新能力〈激進式或漸進式〉，盤點知識基礎〈人力資源及組織資本〉，透過知識網路定位〈相對創業型定位或相對緊密型網路〉與知識策略〈相對知識探索或相對知識運用〉的搭配，提升所需的創新能力。 / Firms aggressively respond to environmental change through innovations. Radicalness of innovations should be guided by environmental demand, or customers’ requirement. When customers’ requirement changes or is unknown, how do firms explore customers’ requirement becomes a crucial question. Contrarily, when customers’ requirement is known or unchanged and competition emphasizes on product/service quality and cost, how do firms exploit knowledge to increase quality or/and to decrease cost becomes a crucial question (Abernathy & Utterback, 1978). How does an innovative firm successfully adopt a proper knowledge networking which acquires above drivers (i.e., new customer requirement, product quality/cost) and facilitates the firm’s knowledge base and knowledge strategy to increase innovation capability? To what extent of knowledge networking, knowledge bases and knowledge strategies combine to affect firm innovation capability? This study tries to answer these questions. According to firms’ responsive degree to environmental change, firms can be categorized into Prospectors, Analyzers, and Defenders from a corporate strategy perspective (Miles et al., 1978). ‘Prospectors’ aggressively respond environmental changes, ‘Defenders’ pursue a stable environment, ‘Analyzers’ attempt to minimize risk while they maximize profit. Corporate strategy directs how does a firm respond environmental change and pursue innovation capabilities (Koka & Prescott, 2008). According to discrepancy degree between market demand and firms’ existing technologies base, innovation capabilities which firms pursue can be categorized into radial innovation capability and incremental innovation capability. Two major schools argue different sources of innovation capability. Knowledge-based scholars (e.g., Subramaniam & Youndt, 2005) argue that innovation capability is from firms’ knowledge (e.g., human capital, organization capital and social capital) utilization and accumulation. Contrarily, knowledge strategy scholars (e.g., Atuahene-Gima, 2005) emphasize importance of adaption. Firms are required to adapting to environmental change, setup radicalness of new product or innovation capability, and deicide resources allocating on exploration projects or/and exploitation projects. Firms increase innovation capability through knowledge exploration strategy and knowledge exploitation strategy. Knowledge-base perspective indicates source of firms’ innovation capability. However, it assumes environment changes slowly or companies are able to dominate environmental change. Contrarily, knowledge strategy perspective indicates direction of firms’ innovation capability for adapting to environment. However, it assumes firms’ knowledge base is fruitful and can be utilized for strategy execution. How does a firm positively interplay knowledge base and knowledge strategy to increase innovation capability is a crucial question (Cook & Brown, 1999). I review literature and firstly indicate that social capital’s characteristic is different from human capital and organization capital. Different degrees of knowledge networking (relatively entrepreneurial positioning and relatively prominent networking) embedding different social capital (new opportunities and trust) facilitates firms’ new opportunities accessibility to increase innovation capability via knowledge base and knowledge strategy. This research deducts eight hypotheses and builds a model which includes innovation capability (dependent variable), knowledge base and knowledge strategy (independent variable), and knowledge networking (moderating variable). Further, I conduct a qualitative research and a quantitative research to test the model. I find Sun-2 workstation in 1980s and Amazon.com website innovation cases from databases and elucidate corresponding variables relationships. Then, I collect items which measure variables from premier journals’ articles to generate a survey questionnaire. I send questionnaires to 341 Taiwan ICT firms’ managers. The firms are listed on Taiwan Stock Exchange. Final 69 effective samples are received (return rate = 20.23%). I use PLS to analyze the data and find that most hypotheses are supported. This quantitative research finds that Taiwan ICT firms emphasize on incremental innovation capability, and incremental innovation projects cannibalize radical innovation projects in human capital, knowledge networking influences innovation capability via knowledge base and knowledge strategy. This research concludes that 1. knowledge networking bridges the gap between knowledge base and knowledge strategy and increases innovation capability, 2. Corresponding knowledge networking, knowledge base, knowledge strategy, and innovation capability on distinct stages of industrial innovation resolves knowledge networking dilemma: A firm which simultaneously adopts relatively entrepreneurial positioning and relatively prominent networking performs worse (e.g., Koka & Prescott, 2008) or better (e.g., Capaldo, 2007), 3. Managers should determine innovation capability portfolio (relatively more radical innovations or relatively more incremental innovations?) by product market’s stages of industrial innovation, check existing knowledge base (human capital and organization capital), adopt corresponding knowledge networking (relatively entrepreneurial positioning or relatively prominent networking), knowledge base and knowledge strategy (relatively exploring knowledge or exploiting knowledge) to increase determinative innovation capability.