Thanks for the invite @Ksenia_C. While I generally support and believe in the need for strong metadata- I’m not well-versed in it outside of what I learned in the one cataloging class I took a very long time ago. I’m super happy to contribute my enthusiastic encouragement, but I’m not sure that I can contribute lots to the central conversation.
1 Like
Thank you for your message Quill 
‘enthusiastic encouragement’ means a lot
Martin Dougiamas’s talk was super interesting, but one of my take homes was that it is a very difficult problem because of the sheer variety of materials to be cataloged.
3 Likes
@Ksenia_C - great topic!
I think as the volume of OE resources and tools grow, there will be an increasing need for meta-data that helps with tasks like categorization, findability, and data exchange. Although I have limited experience in OER meta-data, I do have substantial experience working with data models more generally and would be happy to donate some time to learn more about OE meta-data frameworks and build models that can support the community needs.
@JoshHalpern - I would agree that the large amounts - and perhaps the loosely structured nature - of data make this a challenging task.
My partner @iwonajs is a PhD researcher in AI / Natural Language Processing. Perhaps we can leverage some of our compute resources to aid in this task…
I’m available all week and would be happy to setup a video chat / collaboration session to discuss further!
Cheers,
Shawn
2 Likes
Hello! I would recommend reaching out to Bobby Bothmann, Metadata Librarian at Minnesota State University. He worked on OER Metadata for his capstone project in the SPARC Open Education Leadership Program, which included an overview of current schema and an analysis of best practices among them.
2 Likes
Also, see the session by @vstojano & @samedi, :sync: The “Summer of Open” Policy & Metadata: Mutually-informed Initiatives to Facilitate Sustained OER Collection Development
1 Like
@Shawn @Ksenia_C @samedi If would be interested in joining this conversation if you set up a session! I am located in Utah/MST, United States and generally available this week. 
Thank so much for the suggestion Abbey. I am aware that there is some excellent work on the metadata front being done by SPARC, but it is always good to have a specific point of reference,
thank you!
@Shawn thank you so much for your message, linked and structured data are certainly connected to metadata and AI could play a huge role in discoverability of content.
Personally, I don’t think I know enough about the different metadata schemas and the challenges yet to have a meaningful discussion about ways to support that. I will, however, add your name and your partner’s to my list of people to reach out to in the future. I think it might be a valuable conversation a few months down the road.
Unless of course, @AnitaYT you have a lot of experience with OER metadata and would like to lead this?
Personally, I don’t think I know enough about the different metadata schemas and the challenges yet to have a meaningful discussion about ways to support that. I will, however, add your name and your partner’s to my list of people to reach out to in the future. I think it might be a valuable conversation a few months down the road.
Unless of course, @AnitaYT you have a lot of experience with OER metadata and would like to lead this?
Haha, I know just enough to make a fool of myself. 
I would not be able to lead a scholarly discussion on metadata, data models, AI/ML. I have some practical experience in using HTML tags and alt-text for enhancing SEO and Universal Design, respectively, in digital content — websites, MS Word, PDFs, Pressbooks, etc.
I’ve spent the past few months deep diving into metadata schemas (general and OER-specific) if you’d like some help navigating.
There were some presentations at last week’s Open Ed that cover metatdata and discovery. The video stream links now go to the YouTube videos. I didn’t notice anyone mention these yet, but appologies if I just missed it.
Collaborative Creation of the OER Metadata Rosetta Stone
We need to talk about OER Discovery
I know there are a great many considerations, but I think that it’s really important to make OER discoverable and understandable to commercial search engines. From what I understand from other library-led SEO, structured data, and linked data projects schema.org can be a big part of this and it seems like there is work underway to extend that for the schema type called learning resource.
1 Like
Within our open source project for an OER Search Index https://www.oersi.org we set on schema.org/CreativeWork for import and export of metadata. At the moment we use a german vocabulary for subjects and resource types, but maybe we should switch/map to ISCED as mentioned by MoodleNet.
1 Like
This is a very interesting question! Thank you for bringing it up. We are using LRMI in the Finnish Library of Open Educational Resources. Before choosing what to use, we compared a few of the most used metadata models, but unfortunately we have information on the comparisons only in Finnish. Mainly we chose LRMI because it has possibilities to describe the OERs with curriculum related information and it is not overly complicated. This is important because we do not have experts making the metadata but it is generated by the creators and users of the OER. We made a national profile of LRMI with translations and recommendations for vocabularies. During our work we also tested it with teachers from different educational sectors and we found some needs for improvement that we have since implemented. We also cooperate with libraries and this has brought some further additions. We are happy to share our experinces with implementing LRMI with interested people.
We are currently also testing an AI-solution to further enrich the metadata (it’s called Annif). According to our current test it would enable more metadata, but not necessarily metadata that helps in searching OER. It might come to it that more development is needed for it to be useful for educational resources.
I will also discuss about this briefly in my lighting talk today in a few minutes: :sync: Designing an OER Platform to Support Continuous Learning and All Educational Sectors
1 Like
Thank you for you lightning talk, @annalin. It was very interesting, as we also have similar questions about how to deal with different education levels in germany.
We are also testing Annif at TIB, but not yet in OER context. So I would be really interested in the results of your tests and possible further steps for development.
Sure! We have a meeting on the first results and continuation of steps to improve metadata mid-December. I could send you an email after that to recap on our findings? Just need your email!
University of Jyväskylä has already implemented Annif with their repository for theses’. For them it has apparently been an improvement. Osma wrote about it and other implementations in this publication: https://www.doria.fi/handle/10024/169004
Thanks @annalin - sounds like great work underway!
I wanted to share with you a new tool that I encountered - the R3 Conclave Platform. I’ll do my best to describe it here, though my Partner is really the technical expert in this domain.
Conclave is an interesting platform as it allows for running AI models on private data sources without the need to disclose the data (data is held in a secured memory and only approved code is allowed to execute). This raises interesting opportunities to run AI on HSPII (aka. Highly Sensitive Personally Identifiable Information) data sets, such as Student Information System data, without disclosure of the actual data. My partner @iwonajs is really the technical expert in this area.
I’d be happy to discuss further if you’re interested.
Cheers,
Shawn
Good to see so much interest in OER Metadata! Just a quick note that my colleagues at hbz did some interesting work on this as well. They set up a professional environment for the development of metadata standards based on Github and Respec (see https://github.com/dini-ag-kim/oer-stoeberspecs, in German, but google translate will do the job) as well as a SKOS-based metadata editor: https://skohub.io/.
2 Likes
Here is a link to a talk at DCMI Virtual from my colleague Adrian about how they are developing a common metadata profile based on JSON-LD and schema.org: https://www.youtube.com/watch?v=9UZVHNZ_zLA (slides). See also information on the whole LRMI session at https://www.dublincore.org/conferences/2020/presentations/lrmi_metadata_in_use/. If you want to learn more about SkoHub that is heavily used in creating the metadata profile and controlled vocabularies, you will find plenty of material in the workshop repo my colleagues currently prepare for SWIB20 which will happen next week. https://github.com/skohub-io/swib20-workshop/blob/main/resources/README.md
2 Likes
Thank you for the interesting links and info @Shawn and @trugwaldsaenger! There is so much interesting going on that I seem to be running out of time to get to know everything, but I marked a slot in my calendar later to get to know these initiatives It’s also a good chance to practice my Deutsch! Great to see LRMI in use and also to read about the struggles in implementation. With the Finnish Library of Open Educational Resources we use oai-pmh-protocol to pass our metadata to libraries, so that means we had to express LRMI in xml which had a few problems that we hopefully found an okay solutions to.