Publikationer från Stockholms universitet
Not a member yet
31237 research outputs found
Sort by
Cultural Barriers to Digital Innovation: A systematic literature review
No full textDigital innovation is vital for economic growth, and organizations across various sectors often fail to reach their goal(s) when they do not have sufficient digital innovation. There are various types of barriers that can hinder digital innovation from occurring sufficiently, and one type of barrier for organizations is cultural barriers – barriers that stem from the socially shared knowledge and behavior of the members of an organization. Understanding cultural barriers to a larger degree could improve digital innovation significantly, yielding various positive economic effects, which justifies researching cultural barriers to digital innovation. The problem that this thesis focuses on is the lack of knowledge about cultural barriers to digital innovation, which makes it more difficult to treat such barriers and to improve digital innovation. The research question of this thesis is: What are the cultural barriers to digital innovation? A systematic literature review as the research strategy was chosen to answer the research question. The method for data collection was to use PRISMA, a standardized way of doing systematic literature reviews, and search in the Scopus database. As part of the method for data collection certain articles were included and certain excluded based on explicit criteria. The method for data analysis used was content analysis, which was used to analyze patterns in the gathered data. The research strategy, method for data collection and method for data analysis were chosen because they were deemed the most suitable for the research question, as well as being feasible and ethical. In total 34 cultural barriers to digital innovation were identified carrying out the content analysis. The cultural barriers to digital innovation identified were grouped into the following categories: relations, risk, adaptability, knowledge, miscellaneous values and mentalities, and overarching culture. The relations category has cultural barriers that are about cultural elements that have to do with relations between members in an organization, the overarching culture has sets of cultural elements distributed over the members that make up cultural barriers, and all other are other cultural barriers stemming from the properties of members. The top three most frequently mentioned cultural barriers to digital innovation in order were: a lack of digital culture, insufficient knowledge management and insufficient leadership. The sector with the most cultural barriers to digital innovation found was the non-specific which encompassed many private sector organizations. This thesis can potentially guide various organizations in how they can mitigate cultural barriers to digital innovation, and it can potentially also guide future research about cultural barriers to digital innovation
Händelsedriven Mikrotjänstarktiektur på Industriella Tillverkningsbolag
No full textDen fjärde industriella revolutionen, även kallad Industry 4.0, kräver att företag är anpassningsbara, snabba och datadrivna, vilket ställer höga krav på IT-arkitekturen. En händelsedriven mikrotjänstarkitektur använder händelser för att utlösa och kommunicera mellan isolerade mikrotjänster. En händelse är en förändring i tillstånd eller en uppdatering, såsom skapandet av en ny order, vilket kan initiera uppgifter eller ytterligare händelser i andra mikrotjänster. Denna arkitektur är vanlig hos teknikföretag som Netflix, Spotify och Amazon, men mindre vanlig inom industriföretag. Studien syftar till att undersöka händelsedriven mikrotjänstarkitektur inom industriella tillverkningsföretag genom att besvara frågeställningen: Vilka fördelar och fallgropar finns med att använda en händelsedriven mikrotjänstarkitektur inom industriella tillverkningsföretag, och hur implementeras arkitekturen i praktiken? En fallstudie genomfördes på ett relativt nystartat industriföretag med en IT-arkitektur baserad på mikrotjänster och händelsedriven kommunikation. Fyra djupintervjuer utfördes med anställda med relevant erfarenhet. Data analyserades genom tematisk analys, vilket resulterade i 27 koder under 8 kategorier, grupperade i 3 teman: Praktisk tillämpning, Fördelar och Fallgropar. I fallstudien kategoriseras mikrotjänsterna utifrån funktionalitet i två typer: i) "Substantiv" för att hantera information, och ii) "Verb" för att utföra uppgifter och automation. Resultaten visar att arkitekturen ökar flexibilitet och effektivitet genom att minska beroenden, vilket leder till större möjligheter för automation, iterativ utveckling och teknologisk innovation på både organisations- och tillverkningsnivå. Utmaningarna inkluderar ökad komplexitet, vilket kräver noggrann modellering och kontinuerligt underhåll. Studien bidrar med värdefulla insikter för industriföretag som överväger att implementera en händelsedriven mikrotjänstarkitektur för att möta kraven från Industry 4.0. En begränsning är att studien baseras på ett enskilt företag utan legacy-system, vilket påverkar resultatens generaliserbarhet. Framtida forskning bör fokusera på en mer omfattande undersökning av arkitekturen på tillverkningsnivå, då detta område är mindre utforskat.Industry 4.0 calls companies to be adaptable, fast, and data-driven, placing high demands on IT architecture. An event-driven microservice architecture uses events to trigger and communicate between isolated microservices. An event is a change in state or an update, such as the creation of a new order, which can initiate tasks or further events in other microservices. This architecture is common among tech companies like Netflix, Spotify, and Amazon but is less prevalent within industrial firms. The study aims to investigate event-driven microservice architecture within industrial manufacturing companies by answering the research question: What advantages and pitfalls exist when using an event-driven microservice architecture in industrial manufacturing companies, and how is the architecture implemented in practice? A case study was conducted on a relatively new industrial company with an IT architecture based on microservices and event-driven communication. Four in-depth interviews were conducted with employees who have relevant experience. Data were analyzed through thematic analysis, resulting in 27 codes under 8 categories, grouped into 3 themes: Practical Application, Advantages, and Pitfalls. In the case study, microservices are categorized based on functionality into two types: i) "Nouns" for handling information, and ii) "Verbs" for performing tasks and automation. The results show that the architecture increases flexibility and efficiency by reducing dependencies, leading to greater opportunities for automation, iterative development, and technological innovation at both the organizational and manufacturing levels. Challenges include increased complexity, which requires careful modeling and continuous maintenance. The study provides valuable insights for industrial companies considering implementing an event-driven microservice architecture to meet the demands of Industry 4.0. A limitation is that the study is based on a single company without legacy systems, which affects the generalizability of the results. Future research should focus on a more comprehensive examination of the architecture at the manufacturing level, as this area is less explored
Mellan etik och AI - Etiska riktlinjer för AI-användning inom akademiska institutioner
No full textGenerativ AI, term för datorstyrda processer som skapar, manipulerar eller syntetiserar data, ofta i form av bilder eller text. Användandet av generativ AI har ökat markant inom utbildningssektorn bland studenter vilket har lagt ansvar på institutioner att ta fram nya riktlinjer för att utbildningen ska hålla en viss standard. Studien syftar till att undersöka vilka åtgärder som kan vara relevanta för institutionen av data- och systemvetenskap vid Stockholms universitet att införa för att bevara den akademiska integriteten och förhindra fusk. Genom en kombinerad metod med både kvalitativ och kvantitativ inslag, deltog både lärare och studenter från institutionen för data- och systemvetenskap vid Stockholms universitet i denna studie genom fysiska intervjuer och enkäter. Resultateten kunde påvisa vikten av inkludering och engagemang av studenter i beslut, tydliga etiska riktlinjer och utmaningar som kan kopplas till dessa för att säkerställa att de utvecklas och kvarvaras över tid. Slutsatsen betonar vikten av tydlighet, kommunikation, etiska överväganden och inkludering av studenter för att utforma etiska riktlinjer gällande användning av generativ AI inom utbildningssammanhang för att upprätthålla balans mellan teknologisk innovation och akademisk integritet.Generative AI, a term for computer-controlled processes that create, manipulate, or synthesize data, often in the form of images or text. The use of generative AI has significantly increased within the education sector to maintain a certain standard of education. The study aims to investigate which measures may be relevant for the department of computer and systems sciences at Stockholms University to implement in order to preserve academic integrity and prevent cheating. Through a combined method involving both qualitative and qualitative elements, both teachers and students from the department of computer and systems sciences at Stockholm university participated in this research through physical interviews and surveys. The results were able to demonstrate the importance of including students in decisions, the importance of clear guidelines and the challenges associated with compliance, monitoring, and maintaining these guidelines over time. The conclusion emphasize the need for clarity, communication, ethical considerations and the inclusion of students in designing ethical guidelines for the use of generative AI in education. This in order to maintain a balance between technological innovation and academic integrity
Reinforcement Learning with Constrained Action
No full textReinforcement Learning (RL) is an important technique in machine learning that has gained special significance in solving real-world problems. Typically RL algorithms work under no constraints on the state of the environment or the actions taken by the agent. But to be better useful for real-world scenarios, RL agents need to learn to work under constraints on states and/ or actions. RL algorithms for constraints on states, also known as safety-constrained RL, are widely researched and a lot of techniques have been invented for solving such problems. Most of these techniques involve formulating the problem as a Constrained Markov Decision Process (CMDP) and applying algorithms like Q-learning to solve them. RL problems with constraints on actions, on the other hand, are not widely researched. There exists a limited number of literature on training RL agents under constraints imposed on the actions that the agent can take at each state. Most of the available methods make use of some kind of action mapping technique - a strategy to map the action chosen by the agent, as per its current policy, to the nearest allowed action according to some action constraint. Advantages of such algorithms are that they always make sure that the constraints are not violated. This is particularly helpful in applications where one wrong action by the agent can have a life-threatening or any other critical consequences, medical applications for example. But there are applications where violations of action constraints are not that critical. In such applications, action mapping might restrict the agent's ability to explore the state space effectively. Modelling the action-constrained problem as a CMDP and using Q-learning to solve it can allow a more relaxed exploration for the agent resulting in a more efficient policy. This work tries to validate this hypothesis by applying Q-learning to a social media problem where the agent tries to enhance the engagement rate by performing different actions while trying to not violate a limit imposed on the number of times it can post content during a given interval. Experiments under this work have established that using Q-learning to solve an action-constrained social media problem formulated as a CMDP can be significantly more effective at enhancing the engagement rate. Even though the action mapping method can make sure that the agent never posts more content than the allowed limit, the achieved enhancement rate is much lower, in fact, lower than that achieved by a random agent. The Q-learning agent, on the other hand, can achieve around 16% better engagement rate for a small number of constraint violations
Policy Innovation in E-Government: Assessing Strategies, Impacts, and Best Practices
No full textThis thesis examines the role of policy innovation in advancing e-government within Sweden's public sector, particularly in relation to the European Union's 2030 Digital Compass. The study addresses the challenges of adapting digital governance to a decentralized system while ensuring sustainability, inclusivity, and resilience. Despite Sweden’s established digital maturity, there are still gaps in understanding how to strategically integrate innovative policies to improve service delivery, citizen engagement, and long-term digital resilience. This research employs a qualitative case study methodology, consisting of interviews with representatives from national, regional, and municipal agencies and document analysis of relevant policies and reports. Thematic analysis was utilized to extract patterns and insights, focusing on policy innovation drivers, strategies, and outcomes. The findings indicate that policy innovation fosters collaboration, strengthens data governance, and enhances public service transparency. Citizens-centered policies and inter-agency collaboration play crucial roles in promoting digital inclusivity. However, challenges remain, including interoperability issues, varying levels of digital maturity across municipalities, and limited resources, all hindering cohesive progress. This study highlights the importance of developing adaptive policy frameworks that align EU objectives with Sweden's unique governance structure. This study presents a comprehensive framework (see Table 5-1) and introduces the SMART model, developed by the author, (Table 9-1 in Appendix C) to facilitate the implementation of digitally inclusive, innovative, and resilient e-government solutions. By utilizing best practices and the measurable, systematic approaches outlined in the SMART model, the findings provide actionable recommendations for policymakers. Furthermore, this research contributes to the broader discourse on sustainable e-government transformation by emphasizing the need for policy adaptability, innovation, and digital inclusivity
Scientific Inquiry as a Knowledge-Oriented Learning Objective in Secondary Education : Knowledge Products from Didactic Practice-Based Research
No full textIn this thesis, scientific inquiry is explored as a knowledge-oriented learning objective within lower and upper secondary education. In previous research about scientific inquiry, three key aspects have been emphasised. First, studies have highlighted the importance of students learning to do scientific inquiry (SI), with activities such as formulating researchable questions. Secondly, students must develop an understanding of the nature of scientific inquiry (NOSI), for example, recognising that there is no single ‘scientific method’. Thirdly, students' meaning-making of scientific inquiry is considered important, involving engagement in open inquiries and navigating the contingent pathways they may entail. This thesis focuses on exploring the consequences of treating these three aspects as explicit learning objectives within science classrooms. The theoretical framework of the thesis is grounded in a pragmatic perspective on knowledge development, emphasising knowledge development as holistic, contextual, fallibilistic, contingent, and transactional. Moreover, characteristics of didactic practice-based research are outlined in the thesis. Didactic practice-based research is characterised by a What – the research object being directed at the relationship between student–content–teacher; a How – a symmetrical and complementary collaboration between researchers and teachers; and a Why – ensuring high ecological validity and thus relevance for educational practice. The thesis is based on three sub-studies. In the first sub-study knowledge products generated through didactic practice-based research are explored, resulting in a typology of different forms of knowledge products valuable for teaching practice. The second sub-study focuses on scientific inquiry, identifying key factors that influence students' learning progression, including their opportunity to take epistemic agency. Furthermore, aesthetics is included in the description of how students make meaning of scientific inquiry. The third study examines a specific aspect of scientific inquiry, namely how students formulate researchable questions. Findings suggest that teaching should be designed to enable students to understand this aspect in relation to the entire inquiry process, allowing them to connect their questions to other aspects of inquiry. The thesis highlights the importance of orienting teaching towards authentic scientific practices, which often involve navigating uncertainty and building upon scientific knowledge to explore the unknown. A central conclusion is that students engaging in open-ended investigations require teacher guidance to develop knowledge in relation to scientific inquiry as a knowledge-oriented learning objective.I denna avhandling studeras naturvetenskapligt undersökande som kunskapsmål i undervisningen på högstadiet och gymnasiet, genom didaktisk praktiknära forskning. Med utgångspunkt i tidigare forskning rörande naturvetenskapligt undersökande som kunskapsmål framträder tre viktiga aspekter. För det första menar forskningen att eleverna behöver få utveckla förmågor i naturvetenskapliga undersökningar, vilket på ett övergripande plan innefattar att formulera naturvetenskapligt undersökningsbara frågeställningar, planera för en metod för att undersöka frågeställningen för att sedan genomföra undersökningen och presentera slutsatser. Vidare behöver de få lära sig om naturvetenskapliga undersökningars karaktär. Det innebär exempelvis att förstå att det inte enbart finns en naturvetenskaplig metod. Avslutningsvis behöver eleverna dessutom skapa mening kring naturvetenskapligt undersökande, vilket innefattar att engageras i öppna undersökningar med de kontingenta vägar den undersökande processen kan ta. I avhandlingen studeras konsekvenser av vad som sker i klassrummet då dessa tre aspekter av naturvetenskapligt undersökande behandlas som kunskapsmål. Dessutom förs ett resonemang kring didaktisk praktiknära forsknings karakteristika och huruvida de olika delstudierna i denna avhandling lever upp till denna karakteristik. Fokus ligger på hur denna forskning kan bidra till lärarprofessionens kunskapsbas. Genom didaktisk praktiknära forskning där forskare och lärare samverkar behandlas frågor som är av relevans för lärarprofessionen. Didaktisk praktiknära forskning karakteriseras av ett Vad – forskningsobjektet är riktat mot relationen elev–innehåll–lärare, ett Hur – en symmetrisk och komplementär kollaboration mellan forskare och lärare, samt ett Varför – att didaktik praktiknära forskning har en hög ekologisk validitet, vilket säkerställer relevans för skolverksamheten. Det teoretiska ramverket i avhandlingen utgår från ett pragmatiskt perspektiv på kunskapsutveckling, vilket innefattar en förståelse för både vetenskaplig kunskapsutveckling och elevernas kunskapsutveckling. Genom detta pragmatiska perspektiv betonas kunskapsutveckling som holistisk, kontextuell, fallibilistisk, kontingent och transaktionell. Avhandlingen bygger på tre delstudier. I den första delstudien analyseras kunskapsprodukter som genererats genom didaktisk praktiknära forskning. Resultatet är en typologi över olika former av kunskapsprodukter som är värdefulla för praktiken. I den andra delstudien fokuseras naturvetenskapligt undersökande som kunskapsmål. Här identifieras viktiga faktorer som påverkar elevernas lärandeprogression, inklusive deras möjlighet att ta epistemisk agens. I den tredje delstudien undersöks en specifik aspekt av naturvetenskapligt undersökande, nämligen hur elever formulerar undersökningsbara frågeställningar. Resultatet visar att undervisningen bör designas så att eleverna ges möjlighet att förstå denna aspekt i relation till hela undersökningsprocessen, genom att eleverna ges utrymme att koppla frågeställningen till andra aspekter av undersökandet. I avhandlingen synliggörs vikten av att undervisningen tar riktning mot autentiska naturvetenskapliga praktiker, vilka ofta innefattar att navigera i osäkerhet och att bygga vidare på naturvetenskapliga kunskaper för att utforska något ännu inte känt. En central slutsats är att elever som genomför öppna undersökningar måste få lärares guidning för att utveckla kunskaper såväl i, om som kring naturvetenskapligt undersökande
Experiences of Siblings of Children with Pediatric Acute-Onset Neuropsychiatric Syndrome (PANS)
No full textPediatric acute-onset neuropsychiatric syndrome (PANS) is a relatively new diagnosis, characterized by an abrupt anddramatic onset of obsessive-compulsive disorder together with neuropsychiatric symptoms. This sudden onset ofsymptoms that occurs within 24–48 h presents tremendous challenges for the affected children and their immediate familymembers. Currently, research on the experiences of siblings of children with PANS is lacking. This study exploredsiblings’ perspectives on living with brothers and sisters with PANS, and their perceptions of own ability and availableresources to manage challenging situations. Nine siblings of children with PANS, aged 10–17 years, participated in semistructured interviews. Informed by Transactional Theory, a combined deductive-inductive approach to qualitative contentanalysis was used for analyzing the data. The findings illuminate that PANS has considerable impact on the lives ofhealthy siblings who experienced fear of aggression and unpredictability of their siblings’ behavior; they felt that theirbasic needs as well as social and emotional needs were often neglected. The results also showed that siblings lackedformal knowledge about PANS and had only a limited number of strategies to cope with stressful situations. However,peers in similar situation were a source of emotional support. Practical implications include the need for development ofcommunity-based, family-centered support programs to reduce the impact of PANS on healthy siblings’ well-being.Directions for further research are outlined
Geometrically Aware Markov Chain Monte Carlo
No full textIn this thesis we study the generalisation of the Metropolis Adjusted Langevin Algorithm to the Riemannian manifold of symmetric positive definite matrices P(n). Specifically, an application to hierarchical models that involve the Wishart distribution are considered. A concrete example is given for modelling the rates of synonymous and non-synonymous substitution in a phylogeny. It is proven that a large class of uniformly log-concave posterior densities attain ε-bounded Wasserstein distance from their invariant measures in O(ε−2) iterations of the Riemannian Metropolis-adjusted Langevin Algorithm. It is also shown that common generalisations of the LKJ-distribution neversatisfy a set of sufficient conditions for this bound. Lastly, it is conjectured that certain conditionsfor attaining the iteration complexity bound may be weakened to hold probabilistically only
