📢 #JobOffer at the ACDH!
We are currently seeking a Training Officer (f/m/x, 20-40 hours per week) to become part of our team in Vienna!
You can apply here: https://oeawnr.onlyfy.jobs/job/ju9gmup2
We are looking forward to reading your applications!
fly51fly (@fly51fly)
프리트레이닝에서 습득된 지식이 감독형 파인튜닝으로 어떻게 전달되는지를 '매직 상관관계' 관점에서 분석한 연구입니다. 사전학습의 특성, 데이터 구성 및 파인튜닝 절차가 지식 이전에 미치는 영향에 대한 이론적·실험적 통찰을 제공하며, 모델 개발과 파인튜닝 전략에 시사점을 줍니다.
🤖 Using AI while coding? Think again.
🖥️ A study conducted by Prof. Dr Sven Apel and his team shows that, compared to human partners, developers working with AI assistants approach code less critically and experience less knowledge transfer.
🔗 Read more: https://sic.link/software
#UdS #SaarlandUniversity #SaarlandInformaticsCampus #ArtificialIntelligence #AI #SoftwareEngineering #PairProgramming #KnowledgeTransfer
I'm pretty happy with the how the content of my new video about #PairProgramming and #KnowledgeTransfer turned out, end even the filming and editing is - while not great - OK for me.
Which is a bad omen: The more I'm happy with a video, the worse it performs 🤣
Here it is anyway, in case you are interested: https://videos.devteams.at/w/2BKu8TDU9fT5S3TKn7Tint
Or here, if you prefer YouTube: https://youtu.be/ySRkHi-xw_Q
#GESISblog #blog #knowledgetransfer
New blog post: In the fourth part of our Wi4impact series, Birte Kuhle writes about approaches and challenges in measuring knowledge transfer by linking data:
A Strategic Community #Roadmap for an #Australian #FAIR #Vocabulary Ecosystem
https://doi.org/10.25911/N6K8-F540
Three years ago, I participated in a very engaged workshop at #ANU on #vocabularies for FAIR #data management. It sharpened how I think about vocabularies. I now see them primarily as a #KnowledgeTransfer tool for representing domain expertise in an actionable form. And I think we do a terrible job both at highlighting how critical they are (particularly in an age where trusted expertise is harder to find) and also at making them easier for others to find and reuse.
I picture this scenario. A student is about to start collecting data for their thesis. They need to make choices about what variables to observe or what questions to ask participants, and they need to think about how they want to represent the results to support their analysis. In the ideal case, the actual data collecting effort is about populating an imagined but initially empty data matrix. If they could be assisted to find the best structured and most widely used (in their domain) vocabularies for any categorical values in their data, it would be possible to generate that template matrix with in-built validation tools, etc. The data they finally collect would have most of its metadata already defined and would be properly interoperable with data collected by others in their domain. Meta-analysis would be much simpler.
I am interested in why tools like this don't really exist, or at least why they are not mainstream. I think it's because vocabularies are seen as such an ultra-nerdy subset of the nerdy topic of #metadata rather than presented as an opportunity to stand on the shoulders of others. What can be done to make them more friendly and intuitive for such purposes?
Finally, after way too many struggles, we have a report and recommendations from from that meeting in 2022. I tried to add some of these ideas to the final product as best I could.
Vocabularies serve as essential anchor points for both humans and machines in effective and efficient data processing. Vocabularies include controlled vocabularies, taxonomies, thesauri, ontologies, and metadata schemas, each of which contributes to an ecosystem that encompasses the people, resources, standards, tools, platforms, policies, and practices that make them accessible and useful for researchers. Currently, Australia’s vocabulary ecosystem is fragmented and lacks effective coordination. To address these challenges, a Vocabulary Workshop was held in 2022, sponsored by the Australian Data Archive, the Australian Research Data Commons, and CODATA. From this workshop, a proposal for a strategic roadmap emerged, followed by numerous community consultations conducted between 2022 and 2024. The resulting Strategic Community Roadmap outlines a pathway for Australia to transition from its fragmented landscape to a cohesive and dynamic FAIR Vocabulary Ecosystem. It presents a Vision, Mission, and 57 recommendations categorised into seven key topics, organised around four Strategic Themes. Each recommendation is prioritised by its importance and urgency for implementation. The goal is to promote wider adoption and greater community engagement with machine-actionable vocabularies, emphasising the social and technical support needed to address current data interoperability challenges. This serves as a call to action to maximise the societal, economic, and environmental benefits that can be derived from our national research and data initiatives.
@grumpygamer how about "neither" because it's at best #WastefulComputing and a waste of time and resources...
Handover.ai – Knowledge transfer made easy
#HackerNews #HandoverAI #KnowledgeTransfer #EasyCollaboration #TechInnovation #AItools
ACDH
sayzard
Reddit Tech VN Bot
Saarland Informatics Campus
dtanzer
GESIS
Donald Hobern
Kevin Karhan 
Hacker News
Mathias Bolt Lesniak