Septentrio Academic Publishing
Not a member yet
6415 research outputs found
Sort by
A comparison between Tsetlin machines and deep neural networks in the context of recommendation systems
No full textRecommendation Systems (RSs) are ubiquitous in modern society and are one of the largest points of interaction between humans and AI. Modern RSs are often implemented using deep learning models, which are infamously difficult to interpret. This problem is particularly exasperated in the context of recommendation scenarios, as it erodes the user\u27s trust in the RS. In contrast, the newly introduced Tsetlin Machines (TM) possess some valuable properties due to their inherent interpretability. TMs are still fairly young as a technology. As no RS has been developed for TMs before, it has become necessary to perform some preliminary research regarding the practicality of such a system. In this paper, we develop the first RS based on TMs to evaluate its practicality in this application domain. This paper compares the viability of TMs with other machine learning models prevalent in the field of RS. We train and investigate the performance of the TM compared with a vanilla feed-forward deep learning model. These comparisons are based on model performance, interpretability/explainability, and scalability. Further, we provide some benchmark performance comparisons to similar machine learning solutions relevant to RSs
Signal and Visual Approaches for Parkinson\u27s Disease Detection from Spiral Drawings
No full textThe development of medical decision-support technologies that provide accurate biomarkers to physicians is an important research area. For example, in the case of Parkinson\u27s Disease (PD), the current supervisions of patients become intrusive, occasional, and subjective. However, new technologies such as wearable devices, signal processing, computer vision, and deep learning could offer a non-intrusive, continuous, and objective solution to help physicians with patient monitoring. The Parkinson\u27s Disease Spiral Drawings public dataset was selected to face PD detection in this work by comparing four representation methods of the X, Y, and the pressure time series: signal, visual, hand-crafted, and fusion. The signal approach uses the Fast Fourier Transform of recording windows and a Convolutional Neural Network for modeling; the visual strategy employs visual transformer features from gray-scale images; the hand-crafted technique utilizes statistics calculated from temporal signals, and the fusion combines the information from the previous approaches. In these procedures, a Random Forest classifier was used for PD detection using the attributes extracted from each type of representation. The best results showed an F1 score of 93.33% and 93.06% at the user level using a signal approach with the three signals for the Static Spiral Task and an image-based proposal with X and Y coordinates for the Dynamic Spiral Task, respectively
Leksikalsk variation i de danske ømål
Get PDFInden for dialektforskningen undersøges dialektal variation typisk med udgangspunkt i et begreb, fx et redskab eller et dyr, med det formål at beskrive og forklare den geografiske fordeling af betegnelser for begrebet. Den pilotundersøgelse artiklen gør rede for, afprøver en metode til at undersøge leksikalsk variation mere bredt. Undersøgelsen sammenligner ordforrådet inden for tre centrale arbejdsprocesser i det førindustrielle bondesamfund: kornhøst, smørfremstilling og udkørsel af gødning, for at få indtryk af hvor forskelligartet terminologien er inden for ømålsområdet, og om de almindeligt udbredte ord har særlige karakteristika. Et særligt fokus for artiklen er at undersøge i hvor høj grad Storebælt udgør en terminologisk grænse. Undersøgelsen bygger primært på Ømålsordbogen og dens seddelsamling. Eftersom det onomasiologiske princip har præget både indsamling, organisering af samlingen og redigering er det et velegnet udgangspunkt for begrebsmæssige undersøgelser, men der er dog også en række fejlkilder man må tage højde for ved udnyttelsen af samlingen
To nye dialektordbøger fra Nordvestjylland
Get PDFI afsnit 1 omtales kort traditionen mht. dialektordbøger i Danmark, i afsnit 2 beskrives en mindre, allerede udgivet ordbog over dialekten i Thy, Thy-udtaleordbog (2020, 192 s.), i afsnit 3-4 en større, generel ordbog over thybomålet (ca. 800 s. i A4-format), som forventes klar til udgivelse i 2023. Afsnit 2: Udtaleordbogen gengiver udtaleformer, som de er noteret i dels dialektale digtsamlinger fra 1930’erne og frem, dels dialektale ordlister mv. modtaget senere. Første del af udtaleordbogen er ordnet alfabetisk efter dialektordene, anden del alfabetisk efter de rigsdanske ord, der er brugt som oversættelse af dialektordene. Afsnit 3: Den generelle ordbog bygger på en stor ordsamling (ca. 17.000 kartotekssedler) fra især ét sogn i Nordthy med fonetisk gengivelse af udtalen af de enkelte ord, nuanceret oversættelse til rigsdansk, eksempelsætning(er), faste ordforbindelser, talemåder, oplysninger om sprogbrugen mv. Denne ordsamling blev først indtastet i en database, Thybasen, og derpå via excel-ark overført til word-filer som grundlag for en leksikografisk redaktion af ordene med opdeling i dels homonymer, dels underbetydninger i de enkelte artikler. På flere kartotekssedler er tegnet skitser af de ting, der omtales; de gengives via en professionel tegner som illustrationer i ordbogen. Afsnit 4: Nærmere om behandlingen af centrale fonetiske særtræk (vestjysk stød mv.) i den generelle Thy-ordbog
En stein i lomma kan bidra til trøst: En kvalitativ studie av helsesykepleiers bruk av gjenstander i helsedialoger i skolehelsetjenesten
No full textPublic health nurses aim to prevent illness and promote health based on the Norwegian national guidelines for school health services. The guidelines state that health dialogues should be supportive and contribute to healthy choices for pupils. Most Norwegian teenagers report good physical and mental health. Nevertheless, mental health challenges in adolescence can occur, and at times continue into adulthood. The need for comfort and support will always be there.
The aim of this article is to shed light on how visual methods and concrete objects, or artifacts can contribute to promoting comfort and consolation in public health nurses\u27 school health dialogues.
This article is based on empirical examples from a qualitative doctoral project with 40 public health nurses working in school health services. Relevant empirical examples of the doctoral project were reinterpreted with a focus on comfort. The findings showed that visual methods such as objects can be offered to pupils as tools to support and help them cope when life becomes challenging. Conclusion: A good relationship between nurse and pupil is essential in comforting conversations. Objects can be useful tools in these health promoting dialogues and continue to be a comfort to the young person after they have left the nurses office. Helsesykepleieres mandat er å arbeide helsefremmende og forebyggende basert på retningslinjer for skolehelsetjenesten. Retningslinjene viser at helsedialoger i skolehelsetjenesten kan bidra til sunne helsevalg hos skoleelever. De fleste norske tenåringer rapporterer om god fysisk og psykisk helse. Likevel kan utfordringer tilknyttet ungdomsalderen, være starten på psykiske helseplager inn i voksenlivet. Behovet for trøst og støtte vil alltid være der. Hensikten med artikkelen er å belyse hvordan visuelle metoder og konkrete gjenstander kan bidra til trøst i helsesykepleiers helsedialog i skolehelsetjenesten. Artikkelen tar utgangspunkt i empiriske eksempler fra et kvalitativt doktorgradsprosjekt med 40 helsesykepleiere i skolehelsetjenesten. Relevante empiriske eksempler av doktorgradsprosjektet ble tolket på nytt med søkelys på trøst. Funnene viste at visuelle metoder som gjenstander kan tilbys skolelever for støtte og hjelp, noe som kan bidra til trøst og mestring. Konklusjon: En god relasjon er grunnleggende i helsesykepleiers helsedialog i skolehelsetjenesten. I tillegg kan gjenstander være nyttige verktøy i dialogen og i etterkant av møtet
Bloomsbury Open Collections: A new model for open access monographs
No full textOpen access brings important benefits to scholarly books and their authors. Research shows open access publication can substantially increase and diversify monograph readership, supporting and amplifying a worldwide interest in scholarship in the arts, humanities, and social sciences. But, while open access provides greater equality for readers, the prevailing fee-based models exclude many authors.
To address this challenge, Bloomsbury has this year launched a new model, Bloomsbury Open Collections, that aims to offer open access for books in a way that is sustainable for us as a publisher, but which does not rely on fees paid by individual authors’ institutions and funders. Through this model, we are aiming to make open access publication available to a wider range of scholarly book authors by spreading the cost across multiple organisations, while providing additional benefits to participating libraries. Our hope is to engage a more diverse set of authors, bringing their work to a wider global audience.
In this poster we share details of how our open access monographs pilot works in practice and how we went about developing the model. We provide insights into how feedback from librarians shaped our approach, and how we have sought to incorporate transparency into the model. We also discuss the principles we used to select titles for inclusion in the pilot, and share some early feedback from our author community
Innovation in publishing: Why we need systems for the review and curation of preprints
No full textWatch VIDEO.
The growing number of papers published openly as preprints on bioRxiv and medRxiv during recent years has highlighted how preprints can accelerate and democratise access to research. However, it has also shown the need for systems of review that help readers easily assess the quality of new findings.
A number of groups are now working to openly peer-review preprints, making them more useful for authors, readers, the broader research community and the general public. Some of these models are also taking advantage of the open nature of preprints to enable researchers from groups traditionally underrepresented in science to participate in public review, sharing their expertise and perspectives with authors, readers and others who engage with preprints. These groups include, for example: ASAPbio–SciELO Preprints crowd review, Biophysics Colab, the Novel Coronavirus Research Compendium, and eLife, which officially launched a new model for publishing in January 2023.
eLife’s new model eliminates accept/reject decisions after peer review and focuses instead on providing quality public reviews and assessments. The output is a Reviewed Preprint, which combines the immediacy and openness of preprints with the scrutiny of peer review by experts. This model feeds into eLife’s overarching ‘publish, review, curate’ mission that puts preprints first.
In this presentation, we will highlight the work of eLife in reviewing and curating preprints, and discuss why transforming preprints into reviewed preprints will have significant benefits for scientists and science more broadly
OPERAS Innovation Lab: Supporting Innovative Scholarly Outputs in Social Sciences and Humanities
No full textScholars see innovative means of disseminating their work and data as a chance to improve the process of sharing ideas with different audiences, thanks to technological affordances. They understand innovation either in terms of form (novel means of communicating ideas through different media), or access (opening up outputs and making them easily accessible) (cf. Maryl and Błaszczyńska 2021, p. 34). What we consider a “scholarly text” has thus become understood as an expression that can employ different media and engage creatively with underlying data.
However, engaging with novel forms of communication poses new challenges to scholars, who may lack competencies, know-how, or adequate resources to take full advantage of innovative outputs (Tasovac et al. 2018). This poster will outline the means of support provided by OPERAS Innovation Lab in establishing interdisciplinary, collaborative workflows for supporting innovative outputs in social sciences and humanities (SSH) throughout their lifecycle. The Lab provides guidelines on how to create and sustain FAIR innovative outputs in the SSH.
In the current infrastructure-development project (OPERAS-PLUS) Lab’s website was created (https://lab.operas-eu.org/) and serves as an innovation observatory, collecting and storing data and studies on innovative outputs as well as current research in that field. Apart from the general introduction to the Lab, the poster will showcase three diverse case studies analysed in the ongoing project. They all aim at addressing the actual needs faced by SSH researchers who decide to use innovative forms of disseminating their output:
The novel publication of project outputs: scholarly toolkit;
Management of an interdisciplinary online journal;
Prototyping software services for open science on the example of a recommender system for open access books based on text and data mining (Snijder 2021).
A case study workflow is firstly discussed with the scholars responsible for the project to identify their needs regarding the process and the challenges they are encountering, such as issues of intellectual and technological sustainability or evaluation of non-traditional outputs. Then, in an iterative process, solutions are prototyped, involving various stakeholders, like publishers or e-infrastructures (Ell and Hughes 2013), to forge best practices and provide practical advice. This process is based on a principle of an open collaboration whereby different users will be able to engage with the process and thus the workflow itself will be open for contributions from the wider community through feedback and consultation events, such as this poster presentation
It’s the incentives, stupid!
No full textWatch VIDEO.
From 2004, a new incentives system for scholarly publishing was introduced in Norway. Part of the funding of the HE institutions has since been based on the amount of scholarly publishing produced per institution, as recorded in the Norwegian CRIS system, Cristin. And the incentives worked. Publishing increased more rapidly in Norway than in comparable countries, over the years following 2004 (Aagard et al. 2014).
But the sheer number of journal articles, anthologies and monographs is not necessarily what best brings science forward. And now the Norwegian government has announced that the publishing count will be dropped from the funding scheme (Meld. St. 14). Thus, we suggest that the right move would be to instead direct the incentives to stimulate good, transparent open science research practice.
The advantages of open science, both for further research advancement and the benefit to society, are well documented (Miedema 2022). This includes early and obstacle-free access to research output in terms of articles, peer review reports and commentaries, as well as access to the research processes with preregistration and registered reports, methods and protocols, as well as FAIR research data that enables reproducible research.
So, what is needed is an incentive system that stimulates such research practices. We suggest an incentive system, through the funding schemes of the institutions. This should favour institutions that follow open science practices. The institutions will thus be incentivized to stimulate open science, and identify and remove possible bottlenecks in their organisations that may hamper execution of open science. As employers of researchers as well as support staff, the managers of the institutions will look for ways to optimize local incentives, and also how their organisation may be designed, to best live up to the ideals of open science practices.
Based on the above, we suggest a systematic evaluation of each HE institution, to document how well they practice such openness in their research. To accomplish this a list of criteria that can be easily monitored and objectively evaluated is needed. Several of the open research output parameters should be easy to monitor and measure through a reporting system. The evaluation should also acknowledge that necessary limitations to open processes and outputs exists, and thus avoid biases between institutions due to the nature of their research.
Incentives directed towards the institutional level to stimulate open research practices will motivate the management to stimulate their organisation in the desired direction
BMJ Impact Analytics
No full textResearch that improves patients\u27 lives matters deeply to the funders, institutions and researchers who dedicate their resources to medical and health research. As part of our vision for a healthier world, BMJ engaged with the medical and health research community about the evidence that demonstrates genuine real-world impact and how it can be discovered. In the course of this research, many research managers in health and medicine shared the related challenges that they face with information overload, time, and data quality. It became apparent that AI and Machine Learning can help with finding and tracking vital evidence of real-world impact and also answering questions about research performance and strategies. BMJ began collaborating with Overton to create a prototype that was tested extensively with funders, institutions and researchers.