CoSMO - Common Sensemaking Ontology

CoSMO - Common Sensemaking Ontology

Welcome to CoSMO! CoSMO is an ontology for characterizing social media posts and their relations with the content they cite. CoSMO is closely related to the CiTO ontology which deals with typing of citations in formal academic research (x supports y, x opposes y, etc). CoSMO borrows concepts from CiTO but is different in 2 main ways: (1) CoSMO is oriented towards characterizing the relations between digital media (and not just research papers), and (2) CoSMO deals with the informal knowledge sharing style common to social media networks.
🚧
CoSMO is under active development and is subject to changes.

How to read

As for now, the Ontology describes relations between an “assertion” subject and a URL. Or between the assertion “types”.
In the table below, you may find the description of each relation or type in the prompt field, as we use LLM to give semantic structure to assertions/posts. The field, display_name is the presentation in the application, and label is the way the LLM classifies the relations.
The ontology is used in our Nanopublication template, which the following graph can illustrate:
notion image

Relations

Relations are of the form of 'assertion' relating to 'URL,’ where the URL can be directed to an academic paper, news article, blog post, podcast or video, etc. The assertion is assumed to be created by an ORCID account expressing.
🎄
SenseNets Ontology
name
display_name
prompt
URI
🗜️ summarizes
this post contains a summary of new research. The summary is likely provided by the authors but may be a third party. We use a broad definition of research that includes classic and non-traditional outputs. Classic outputs include papers, datasets or code. Non traditional outputs can include a podcast, blog post, video explainers, etc.
📜 mentions-call-for-papers
this post contains a call for research papers, for example to a journal, conference or workshop.
➕ endorses
this post endorses the mentioned reference. This label can also be used for cases of implicit recommendation, where the author is expressing enjoyment of some content but not explicitly recommending it.
👎 disagrees-with
this post disputes or expresses disagreement with statements, ideas or conclusions presented in the mentioned reference.
👍 agrees-with
this post expresses agreement with statements, ideas or conclusions presented in the mentioned reference.
👀 indicates-interest
this post indicates interest in a reference. This label is meant for cases where the post is not explicitly recommending or endorsing the cited reference.
🏦 mentions-funding
this post mentions a funding opportunity, for example a research grant or prize.
📺 watching-status
this post describes the watching status of the author in relation to a reference, such as a video or movie. The author may have watched the content in the past, is watching the content in the present, or is looking forward to watching the content in the future.
📑 reading-status
this post describes the reading status of the author in relation to a reference, such as a book or article. The author may either have read the reference in the past, is reading the reference in the present, or is looking forward to reading the reference in the future.
🎧 listening-status
this post describes the listening status of the author in relation to a reference, such as a podcast or radio station. The author may have listened to the content in the past, is listening to the content in the present, or is looking forward to listening the content in the future.
🔗 links-to
This is a special tag. Use this tag if none of the tags above are suitable. If you tag a post with <default>, no other tag should be assigned to the post.
🧐 reviews
this post contains a review of another reference, such as a book, article or movie. The review could be positive or negative. A review can be detailed or a simple short endorsement.
👌 recommends
The author is recommending any kind of content: an article, a movie, podcast, book, another post, etc.
❔ ask-question-about
this post is raising a question or questions about some content it's referring to. The content could be a research paper or other media like a podcast, video or blog post.
📝 quotes-from
this post is quoting text from an article it's referring to. Symbols like ">" or quotation marks are often used to indicate quotations.
🗣️ discusses
this post discusses how the cited reference relates to other facts or claims. For example, post might discuss how the cited reference informs questions, provides evidence, or supports or opposes claims.
🗓️ announces-event
this post includes an invitation to an event, either a real-world or an online event. Any kind of event is relevant, some examples of such events could be seminars, meetups, or hackathons. This tag shold only be used for invitations to events, not for posts describing other kinds of events.
📢 announces-job
this post describes a job listing, for example a call for graduate students or faculty applications.
📢 announces
this post contains an announcement of new research. The announcement is likely made by the authors but may be a third party. We use a broad definition of research that includes classic and non-traditional outputs. Classic outputs include papers, datasets or code. Non traditional outputs can include a podcast, blog post, video explainers, etc.

Types and Classes

We define a class - semanticPost, all assertions graphs in our applications will be of type semanticPost. We define a few other types that are assigned to assertions without URL as relation objects. In RDF, we realize these by subclasses of assertion. E.g.
this:assertion rdf:type cosmo:quesion
🎄
SenseNets Ontology
name
URI
display_name
prompt
notes
⬛ possible-missing-reference
this post seems to be referring to a reference by name but has not explicitly provided a URL link to the reference. For example, a post that discusses a book and mentions it by title, but contains no link to the book.
🔭 discourse-graph/observation
this post is articulating a single, highly observation. The intuition is that observation notes should be as close to “the data” as possible. They should be similar to how results are described in results sections of academic publications.
🫴 discourse-graph/claim
this post is articulating an idea or a claim
❓ discourse-graph/question
this post is raising a research question.
semanticPost, all assertions graphs in our applications will be of type semanticPost. This is a placeholder.

Literal relations

Name
URI
Description
a relation between the post/assertion and a list of (literals) words that are found to be descriptive of the post domain, intensions and main argument.

Technical relations

Technical relations
Name
Description
URI
Range
Domain
A relation that indicates that a nanopub is signed with an RSA key that is generated by the object hashed key. For example, in our app we use an Ethereum wallet address.
This is an intentioned subclass of http://www.w3.org/ns/prov#Activity which represent the process of an agent creating a semantic post with an NLP model assistant. Note that the model does not determine the outcome but only suggest structure.
This relation expresses that the entity was automatically published by an software agent, and was not reviewed by a human one.
A special relation for quoting tweets/toots and etc…
post URL
assetions

Nanopublications instances and Templates

Nanopub templates