Search

I verkliga beslutsanalyser, oavsett om det är i mänskliga övervägda situationer eller icke-mänskliga situationer (semi-intelligent maskin/agent i realtid), där är väldokumenterade problem när det gäller framkallande av sannolikheter, verktyg, och kriterievikter. I detta dokument undersöker vi automatisk viktgenerering med flera kriterier metoder med en detaljerad undersökningsmetod som inte setts tidigare. De resultaten bekräftar att Sum Rank-metoden för ordinalfallet och motsvarande Cardinal Sum Rank-metoden för kardinalfallet, överträffar alla andra metoder för robusthet. Nya fynd inkluderar att det inte finns någon indikation att skillnaden i resultat i viktgenereringen minskar eftersom antalet frihetsgrader växer vilket tidigare ansågs vara sant. Som väntat överträffar kardinalmodellerna vidare de ordinarie modellerna. Mer oväntat, men prestandan för den dominans intensitetsbaserade vikten modeller är på sin höjd mediokra för vissa kombinationer och inte ens lämplig för andra kombinationer. En annan insikt från utredningen i denna tidning är den tidigare litteraturen är inte homogen i modelleringen av attributvärdena, vilket resulterar i inte alla metoder som beaktas i denna undersökning kan direkt jämföras., In real-life decision analysis, whether in human-deliberated situations or non-human (real-time semi-intelligent machine/agent) situations, there are well-documented problems regarding the elicitation of probabilities, utilities, and criteria weights. In this paper, we investigate automatic multi-criteria weightgenerating methods with a detailed investigation method not seen before. The results confirm that the Sum Rank method for the ordinal case, and the corresponding Cardinal Sum Rank method for the cardinal case, outperform all other methods regarding robustness. New findings include that there is no indication that the difference in the results in the weight generation is diminished as the number of degrees of freedom grows which was previously thought to be true. Further, as expected the cardinal models outperform the ordinal models. More unexpectedly, though, the performance of the dominance intensity-based weight models is at most mediocre for some combinations and not even suitable for other combinations. Another insight from the investigation in this paper is that previous literature is not homogeneous in the modelling of the attribute values, resulting in not all methods considered in this investigation can be directly compared.

As artificial intelligence systems permeate society, it becomes clear that aligning the behaviour of these systems with the values of those involved and affected by them is needed. The value alignment problem is widely recognised yet needs addressing in a principled way. This paper investigates how such a principled approach regarding online institutions a class of multiagent systems can provide key insights on how the value alignment problem can be addressed in general.

As artificial intelligence systems permeate society, it becomes clear that aligning the behaviour of these systems with the values of those involved and affected by them is needed. The value alignment problem is widely recognised yet needs addressing in a principled way. This paper investigates how such a principled approach regarding online institutions a class of multiagent systems can provide key insights on how the value alignment problem can be addressed in general.

Collections of time series can be grouped over time both globally, over their whole time span, as well as locally, over several common time ranges, depending on the similarity patterns they share. In addition, local groupings can be persistent over time, defining associations of local groupings. In this paper, we introduce Z-Grouping, a novel framework for finding local groupings and their associations. Our solution converts time series to a set of event label channels by applying a temporal abstraction function and finds local groupings of maximized time span and time series instance members. A grouping-instance matrix structure is also exploited to detect associations of contiguous local groupings sharing common member instances. Finally, the validity of each local grouping is assessed against predefined global groupings. We demonstrate the ability of Z-Grouping to find local groupings without size constraints on time ranges on a synthetic dataset, three real-world datasets, and 128 UCR datasets, against four competitors.

Cloud computing is an emerging technology where organizations can have flexibility of their infrastructure and can store and retrieve data from cloud storage. The main advantages of cloud computing are speed, cost reduction and scalability. Cloud computing is mainly developed on virtualization of re-sources with reusability capabilities. In virtualization computer systems are deployed virtually as software and hardware resources are shared between different virtual machines. It allows multiple operating systems and software applications run on a single server. Virtualization plays an important role in the cloud environment. Though cloud computing improves the performance and quality of work for organizations it has security, privacy and trust related threads. In this paper we have surveyed the security aspects of virtualization. We have discussed the cloud computing and virtualization followed by de-tailed analysis of virtualization security.

This paper presents a study of retellings about Genshin Impact, exploring how the game’s narrative design inspires players to create fiction and art based on the game’s universe. A questionnaire sent to players rendered 1606 replies. Based on the findings in this corpus, eight players and creators of retellings were interviewed in-depth. Among our findings were that players were most inspired by the characters, detailed worldbuilding, and regional cultures in the game world. Their motivation to create was often spurred by the gaps and ambiguities in the detailed narrative design, wanting to “fill in the gaps” and, through their creation of fiction, further explore the intricacies of the game’s narrative elements.

In reinforcement learning, an agent makes decisions to maximize rewards in an environment. Rewards are an integral part of the reinforcement learning as they guide the agent towards its learning objective. However, having consistent rewards can be infeasible in certain scenarios, due to either cost, the nature of the problem or other constraints. In this paper, we investigate the problem of delayed, aggregated, and anonymous rewards. We propose and analyze two strategies for conducting policy evaluation under cumulative periodic rewards, and study them by making use of simulation environments. Our findings indicate that both strategies can achieve similar sample efficiency as when we have consistent rewards.

Educational concepts are the core of teaching and learning. From the perspective of educational technology, concepts are essential meta-data, represen- tative terms that can connect different learning materials, and are the foundation for many downstream tasks. Some studies on automatic concept extraction have been conducted, but there are no studies looking at the K-12 level and focused on the Swedish language. In this paper, we use a state-of-the-art Swedish BERT model to build an automatic concept extractor for the Biology subject using fine- annotated digital textbook data that cover all content for K-12. The model gives a recall measure of 72% and has the potential to be used in real-world settings for use cases that require high recall. Meanwhile, we investigate how input data fea- tures influence model performance and provide guidance on how to effectively use text data to achieve the optimal results when building a named entity recognition (NER) model.

Epidemics of infectious diseases can pose a serious threat to public health and the global economy. Despite scientific advances, containment and mitigation of infectious diseases remain a challenging task. In this paper, we investigate the potential of reinforcement learning as a decision making tool for epidemic control by constructing a deep Reinforcement Learning simulator, called EpidRLearn, composed of a contact-based, age-structured extension of the SEIR compartmental model, referred to as C-SEIR. We evaluate EpidRLearn by comparing the learned policies to two deterministic policy baselines. We further assess our reward function by integrating an alternative reward into our deep RL model. The experimental evaluation indicates that deep reinforcement learning has the potential of learning useful policies under complex epidemiological models and large state spaces for the mitigation of infectious diseases, with a focus on COVID-19.

This research is an interdisciplinary work between data scientists, innovation management researchers and experts from Swedish academia and a hygiene and health company. Based on this collaboration, we have developed a novel package for automatic idea detection with the motivation of controlling and preventing healthcare-associated infections (HAI). The principal idea of this study is to use machine learning methods to extract informative ideas from social media to assist healthcare professionals in reducing the rate of HAI. Therefore, the proposed package offers a corpus of data collected from Twitter, associated expert-created labels, and software implementation of an annotation framework based on the Active Learning paradigm. We employed Transfer Learning and built a two-step deep neural network model that incrementally extracts the semantic representation of the collected text data using the BERTweet language model in the first step and classifies these representations as informative or non-informative using a multi-layer perception (MLP) in the second step. The package is called AID4HAI (Automatic Idea Detection for controlling and preventing Healthcare-Associated Infections) and is made fully available (software code and the collected data) through a public GitHub repository. We believe that sharing our ideas and releasing these ready-to-use tools contributes to the development of the field and inspires future research.

What happens when scholars approach novel phenomena such as Interactive Digital Narrative (IDN)? How can we be certain that theoretical frameworks, analytical approaches, and vocabulary are adequate,which means that they are able to fully describe the specific characteristics of the novel phenomena? The same goes for approaches in the practice - how can we be sure that the chosen design methods enable the use of the full expressive potential of a novel phenomena? Furthermore, we ask how we can critique and improve categories and approaches? We reflect on how theory and analytical approaches have been produced so far, identify issues with the current practice, consider alternatives and propose a number of measures to improve the situation. Amongst them are increased efforts on the meta-level in terms of theoretical development and reflective works which concern themselves with the further development of the field, iterative approaches towards theory and method, a more critical approach in education, multi-method analysis, and dynamic representations.

Automated Essay Scoring (AES) automatically assigns scores to essays at scale and may help to support teachers' grading activities. Recently, AES methods based on deep neural networks (DNN) have significantly improved upon the state-of-the-art performance by learning relations between holistic essay scores and student essays. However, DNN-based AES methods function like black-box, negatively affecting the ability to provide automated writing evaluation (AWE). In this work, we proposed a new method, topic-aware BERT, based on fine-tuning the pre-trained language model to learn relations between essay scores and text representations of student essays as well as topical information in essay writing instructions. Moreover, we propose an approach to automatically retrieve key topical sentences in student essays by probing self-attention maps in intermediate layers of topic-aware BERT. We evaluate the performance of topic-aware BERT to (i) perform AES and (ii) retrieve key topical sentences using the open dataset Automated Student Assessment Prize and a manually annotated dataset, respectively. Our model achieves a strong AES performance compared with previous state-of-the-art DNN-based methods and shows effectiveness in identifying key topical sentences in argumentative essays.

A pattern is a concept widely used in engineering and other disciplines for variety of purpose, e.g. business or system analysis. In this work, we present the concept of patterns usage within enterprise modelling. The undertaking is a part of the larger case study dedicated to Fractal Enterprise Modelling (FEM) development deploying Design Science (DS) research methodology. Based on the case study, we have identified three FEM related patterns: one modelling pattern and two analyzing patterns. A modeling pattern advises on a proper way of building a model for a particular purpose, while an analysis pattern helps to find places in a business where a decision could/should be made. More precisely, the modelling pattern identified in this study helps to build a model on an appropriate level of granularity for a certain purpose (operational decision-making). The analysis patterns identified in the study are divided into two categories of patterns: transformational patterns and problem patterns. Transformational pattern is a pattern that contains a condition and an action parts that represent a standardized solution or an opportunity. Problem pattern is a pattern that expresses conditions where a problem might exist without specifying a definite action. These patterns contribute into the creation of the bank of patterns to guide practitioners in modelling and analyzes of the business situations using FEM.