Insa Schaffernak

🌮 Food for thought 🧠 #PublicHealth is complex. This has already been discussed in plenty of papers. Consequently, the use of #digital tools to achieve typical Public Health Goals (aka #DigitalPublicHealth) could not be less complex. Right? Together with my colleagues Hans-Henrik Dassow, Daniel Diethei, Merle Freye, Jasmin Niess & Stefanie Do, I worked on mapping the fields related to Digital Public Health. The result is this colourful sunburst diagram ☀️ 🔎 What did we find? 🔍 1) Digital Public Health (DiPH) - as for our purpose defined as the overall field - includes the clinical & individual-centered #DigitalHealth 🎯 2) DiPH is colourful! It consists of the overarching fields: Humanities, Environmental Sciences, Social Sciences, Engineering & the Natural Sciences. These can be split into individual disciplines, which also come with their subfields (we kept the figure simple and did not portray every subdiscipline here) 🖍️ 3) Due to this, DiPH is extremely #interdisciplinary - but also the individual subdisciplines are interdisciplinary (especially in the Environmental Sciences). 💫 💥 What does this mean for public health practice?💥 1) The interdisciplinarity of DiPH-related projects should be reflected in the aim, methodology, and selection of team members 🪞 2) Multidisciplinary & interdisciplinary teams may use different language & methods --> be inclusive and communicate clearly! 💬 3) Allow for (in-)formal exchange between team members. Different approaches spark interest. Talk about this! 🥂 4) Be open towards different approaches from another field of research or practice. Your gold-standard methodology might not be the best fit for a specific aspect of the project 🤗 5) Be patient! All these different approaches combined in one project will take longer than projects from one discipline. BUT they have the power to produce more robust & holistic results 🦾 📚 Read more on our thoughts in our open-access handbook here: https://lnkd.in/eeaFS83V

316

10 Kommentare

Dennis Dobler Appointed Professor of Computational Statistics at RWTH How effective is a new medication? When do side effects occur? And how should researchers deal with missing information in clinical studies? These are among the questions that drive the work of Professor Dennis Dobler an expert in survival analysis who joined RWTH’s Faculty of Mathematics, Computer Science and Natural Sciences as professor of Computational Statistics in June. Survival analysis is a branch of statistics that examines the time until a particular event occurs – for example, the onset of side effects, disease relapse, or patient death. “These data sets are often incomplete, because for many patients the event of interest doesn’t occur during the study period,” Dobler explains. “That makes developing analytic methods both challenging and exciting.” 💯 Read more ➡️ https://lnkd.in/eyqq9axT 📸: Dobler

142

3 Kommentare

In the news: I did not make it to the hairdresser before, but thank you to Christian Feld and Tagesschau for still featuring me on the German evening news for the start of the AI Act general-purpose AI model rules. We discussed a number of things; the quote that ended up on TV is about copyright transparency rules on how they might enable follow-on damage claims by rights holders. It is a bit hard, however, to do justice to a highly complex regulatory regime in 15 second statements :-) So, here's my more extended take: 1) I am quite happy about: – the prohibitions (Article 5) – AI literacy (Article 4) – general transparency (Article 50) – the GPAI rules (Articles 51-56) See: https://lnkd.in/enjKwNT2 2) I have more mixed feelings about the high-risk rules, particularly due to unresolved interactions with sectoral regulations, the GDPR, and tort law. See: https://lnkd.in/e6UAd_ai 3) The simplification process will hopefully be a good point to clean this up without gutting the core parts of the Act (see 1)). 4) The concept of "systemic risk" remains an epic fail due to the unnecessary restriction to the "most advanced" models. Hallucinations – a key problem with Generative AI – are likely not covered, for example. Paper on this forthcoming. 5) Regulation is, of course, not generally bad for innovation. Empirical studies show that it very much depends on the type of regulation, and that legal regimes can actually foster innovation. Europe's problems with AI advancement have little to do with the AI Act and a lot with lack of VC financing, access to a absolute top-notch talent, and the general risk mentality among significant parts of the population and incumbent entrepreneurs. See: https://lnkd.in/e43tuk7J; https://lnkd.in/emQEvHFJ; https://lnkd.in/emvfVY59 6) Now: off to the mountains for some decent digital detox. Enjoy your summer! European New School of Digital Studies Europa-Universität Viadrina Frankfurt (Oder)

349

18 Kommentare

After 5-years of the German National Research Data Initiative NFDI, there are currently big discussions on how to continue. The German Science and Humanities Council presented an evaluation report: https://lnkd.in/eyZWt7fS The main recommendation here is to reform the governance structures and provide long-term funding. While I agree with the latter, changing/expanding the governance structures will not improve the situation, but rather make things worse - planning economy-style committees will never be able to make the decisions to create user-friendly, attractive services. We need to put the services at the center of attention much more. Since I have repeated deja vus with such initiatives - we seem to make the same mistakes over and over again - I created this analysis and summary of my thoughts on the topic: https://lnkd.in/eeWXtNHY In essence, we need to fund protagonists developing services instead of consortia, focus on a single, clearly defined KPI (usage), have an incremental/agile maturity-oriented development model, and reduce governance to an absolute minimum (essentially evaluating service proposals and success). I have now added this slide summarizing what I think we should do to make the NFDI a success:

92

12 Kommentare

#ResearchAlert - Digitale Transformation 🤖 1. Sandra M. Siedl and Martina Mara (LIT Robopsychology Lab) about wearing a work exoskeleton and its effects on users' perceptions of their own competence, attractiveness, and humanity. 👉 http://go.jku.at/w25e6 --- 2. Vasiliki LAINA, PhD (JKU School of Education, University of California, Berkeley​) about the use of digital tools in secodary education and their collaborative character 👉 http://go.jku.at/w25f7 --- 3. Florian Schmid, Khaled Koutini and Gerhard Widmer (JKU - Institute of Computational Perception​ & LIT AI Lab) about dynamic CNNs and if they outperform Transfoers in audio tasks with less computation. 👉 http://go.jku.at/w25f4 #DigitaleTransformation #jkulinz

60

1 Kommentar

𝗚𝗿𝗲𝗲𝗻𝗹𝘆𝘁𝗲 𝗿𝗲𝗰𝗲𝗶𝘃𝗲𝗱 𝗳𝗶𝗻𝗮𝗹 𝗮𝗽𝗽𝗿𝗼𝘃𝗮𝗹 𝗳𝗼𝗿 𝗺𝘂𝗹𝘁𝗶-𝗺𝗶𝗹𝗹𝗶𝗼𝗻 𝗳𝘂𝗻𝗱𝗶𝗻𝗴 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱 𝗟𝗶𝗾𝘂𝗶𝗱𝗦𝗼𝗹𝗮𝗿 𝗠𝗲𝘁𝗵𝗮𝗻𝗼𝗹. Yesterday, Dr. Korinna Kunde and 🌱🌎 Florian Hildebrand accepted the official grant notice of approval from minister Mona Neubaur - confirming multi-million euro public funding co-financed by the Ministerium für Wirtschaft, Industrie, Klimaschutz und Energie des Landes NRW and  EFRE/JTF in NRW by the European Union through the Produktives.NRW program. 𝗧𝗵𝗶𝘀 𝗴𝗿𝗮𝗻𝘁 𝗲𝗻𝗮𝗯𝗹𝗲𝘀 𝘁𝗵𝗲 𝗹𝗮𝘂𝗻𝗰𝗵 𝗼𝗳 𝗼𝘂𝗿 𝗙𝗢𝗔𝗞 𝗳𝗮𝗰𝗶𝗹𝗶𝘁𝘆 𝗟𝗶𝗾𝘂𝗶𝗱𝗦𝗼𝗹𝗮𝗿 𝗠𝗲𝘁𝗵𝗮𝗻𝗼𝗹, 𝗯𝘂𝗶𝗹𝘁 𝗶𝗻 𝗰𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 Evonik 𝗮𝘁 𝘁𝗵𝗲𝗶𝗿 𝗰𝗵𝗲𝗺𝗶𝗰𝗮𝗹 𝗽𝗮𝗿𝗸 𝗶𝗻 𝗠𝗮𝗿𝗹.  The plant will integrate Greenlyte’s proprietary carbon capture and hydrogen technology with methanol synthesis at a meaningful size.  It marks our transition from pilot to commercial viability and serves as a lighthouse project: the first to combine carbon capture, green hydrogen production and e-fuel synthesis in one integrated system at this scale. “We’re incredibly grateful for the support we’re receiving in NRW,” says 🌱🌎 Florian. “This is a place where legacy industry, innovation, and strong academia actually come together - while elsewhere it often stays at the talking stage. NRW shows that collaboration between these stakeholders can lead to big success and create actual value, both economically and ecologically.” Thank you to everyone who made this project and funding program possible. 💚 Stay tuned for the next updates! 🔗 Check out the link in the comments to find out more about the project. Dr. Niklas Friederichsen 🌱🌎 Florian Hildebrand Dr. Martin Schmickler Vincent Daub Dr. Korinna Kunde

327

19 Kommentare