Cohort, mentors, friends: Is anyone interested in discussing capstone projects here?
My own capstone ideas
I could go either way, based on the capstone guideline: Either work on a KnowHax project (DREAM-KG) or my own project. Following is a copy of a message I posted in an Explorer cohort discussion:
I’m working on a portfolio project to build a tool for curating and sharing bookmarks and citations from articles found on the Web: LinkedIn, blogs, Twitter, any web site. This is in the field of personal content and shared knowledge management.
Using this tool, individuals will capture bookmarks with metadata comparable to scholarly citations in JSON-LD format. Articles usually contain links, and the user will have the option to scrape the content and choose among links. These links would in turn be captured in a standard citation format.
Graph data is the natural structure for a semantic layer for this applications. I would create an ontology that captures entities for articles, authors, users, and bibliographic information. Fortunately, ontologies exist, and I would include them, rather than reinventing. (FOAF, ORKG?)
AI/NLP comes into play for (local / personal) analysis and capture of information contained in cited material. One of the challenges would be to make this a) fast and b) feasible without access fees.
Nice-to-have: a chatbot to interrogate the materials captured, and to help capture the personal information that would help others use the citations.
As for the KG itself, I’d prefer to keep all data local, but there are only a few technologies that can be deployed so easily. For version 1 I would run a database (eg supabase) in the cloud. This would create challenges around security and privacy. I’ll watch Samsung carefully to see if and how they deploy personal KG tech based on their acquisition of RDFox.
1 Like
Thank you for this KEith. Would to like to connect to see how we can articulate / align a challenge at KnowHax in March, around your idea above?
(please send an email to hugo@genhax.com for faster answers)
I love the idea of the citations project, Keith. I am still unsure of how this capstone project is supposed to work though.
Hi All,
i posted in the builder classroom assignment but am reposting here in case others have not chosen capstone projects yet.
Is there anyone else here who has not yet chosen a project and has images in a domain in which they are subject matter experts and/or images they would like to work with using ImageSnippets? (https://imagesnippets.com/).
I was reading back through the projects and I just sent Maru a very long personal message on this. I know it is past the deadline for ‘choosing’ but I have submitted a choice for 2 different things that I believe I could work on simultaneously.
Is there anyone who has not selected their capstone project yet? One of the project I could work on, would be for me work with anyone here who may have a small collection of images, probably between 15-25 images for the capstone which would actually be a LOT of material to consider and work on for domain modeling/knowledge graph building in an interesting domain that they would like work with?
It could be any kind of domain as a use case, preferably one where someone has some subject matter expertise. The use case/project would be around what all would be involved (many hands on topics/lessons) in using the images for engineering an ontology using images.
For this project, I would not only explain all of the considerations that go into building knowledge graphs from image descriptions, but would also work with the domain (which is where I would get my own fresh hands-on work) — i.e. the person bringing the images might be a subject matter expert with their images in a domain I don’t know much about and together, we would have to make a lot of considerations about how we would want to construct the ontology in this domain. If anyone would like to approach me about this - please reach out!
As my other project, which I want to do anyway is to work with Keith on the citations project. Last week, Keith and I both were introduced to an architecture that neither of us know much about and would like to learn more about. We both immediately thought it would be ideal for working on the citations/bookmarking project. Additionally, we might also consider how the ideas for the triple-editor which is a part of ImageSnippets could also be used in association with the Virtuoso server/Opal architecture we were exploring last week.