spothedog1

1 will not be standardized deterministically, 2 will require a lot of work from me to figure out how to structure an organization. There are a million edge cases to think through

I think a big value add would be to just automatically use the Organization Ontology [1] and everyone who uses Graphiti on a text block that describes an organization will get that back.

Then defacto everyone's schemas can interact with each other. My business value prop is not going to be in defining an organization's structure, but it's still useful data for me have. Forming good Ontologies takes a lot of work.

I was thinking of just fine tuning my own LLM with RDF ontologies so it always returns graphs formatted in commonly used RDF schemas

[1] https://www.w3.org/TR/vocab-org/

EDIT: Just saw your edit 2 comments above. And yes exactly that is what I think would be useful. I'm looking for a project in this space that is opinionated cause frankly using a graph with a random assortment of node and edge types everytime I run it is not very useful.

Comment · 2 years ago

Can you explain why you think Dublin Core is a bad standard?

Comment · 2 years ago

Very cool, I've been working on getting Qlever setup and using it as my main triple store. Was excited to see this hit my feed

Comment · 2 years ago

The reason I bring RDF is because I use ontologies that have been defined by experts and covers ton of edge cases. If a group of genealogists define a `fam:` RDF Ontology and publish it, then I want every family relationship in my graph to use their Ontology.

I'm looking for something like graphiti that can take in a text block and when creating the relationships, automatically know to use the `fam:` ontology when creating familial relationships. The vast majority of people don't feel like defining schemas for every little thing and they're basically the same across all systems except for custom proprietary ones you define as your IP.

Their ontology would have OWL rules like `fam:isChildOf` `owl:inverseOf` `fam:isParentOf` so running an OWL Reasoner over the graph would generate the inverse triples as well

So if I had the text `Joe is Bob's dad`, input it into graphiti, then get the triples

person:Joe fam:isParentOf person:Bob person:Bob fam:isChildOf person:Joe

and the edge would be in a shared definition amongst all graphiti users. The LLM can be fine tuned to recognize exactly what fam:isParentOf means so there is no ambiguity. Right now I'm guessing graphiti could spit out edges `IS_SIBLING_OF` `SIBLING` `SISTER` `BROTHER` etc, its not standardized which makes it difficult to interact with computationally if say, I wanted to input a bunch of random text and then run pre-trained graph models of family networks.

Comment · 2 years ago

Looks cool, would love support for RDF Graphs. The reason I prefer those is because the ontology is already well defined in a lot of cases which is 80% of the battle with Knowledge Graphs in my experience. Without a well defined Ontology I think LLM <> KG integration will not live up to its potential. LLMs have to know what nodes and edges really mean across diverse datasets

Comment · 2 years ago

Following up: A reason I really like using Jupyter in PyCharm is because Github CoPilot works in it which helps a lot.

Comment · 2 years ago

I'm very interested in this. I'm a Software Engineer who's been doing some Data Science on the side and been looking for something like this.

My current set up is running Jupyter on an EC2 instance and using inside PyCharm. One feature I actually really value is being able to use it directly in PyCharm as I can have my IDE on one side of split screen and my browser on the other. Not sure how feasible it is to integrate something like this into an IDE, VSCode would work

But a real killer feature that could get me to switch to a browser based would be the ability to load custom context about the data I'm working with. So I have all my datasets and descriptions of all their columns in my own database and would love a way to load that into the LLM so that it has a greater understanding of the data I'm working with in the notebook.

I store all my data in objects called `distributions` [1] and have a `get_context()` function that will return a text blob of things like dataset description, column description, types, etc.

The issue with all these auto-code AI tools is they don't really have a good grasp of the actual data domain and I want to inject my pre-made context into an LLM thats also integrated in my notebook.

[1] https://www.w3.org/TR/vocab-dcat-3/#version-history

Comment · 2 years ago

twitter only shows the first tweet if you're not logged in, can you post the repo here

Comment · 2 years ago

Can you expand on Postgres vs Graph Databases?

Comment · 2 years ago

This is what I've been thinking for the past year. Do you have any more resources where I can explore what you're talking about?

* What are some deductive reasoning engines out there?

* How are you storing the rules in a graph? Using triples?

* What query languages are you using for this?

Comment · 2 years ago

My personal exploration has me moving in the direction of using LLMs as a Language Model and some sort of structured knowledge representation as a Knowledge Model. So using knowledge graphs to store facts and data and LLMs to interface with it. I wish there was more focus on this, right now I'm just using RDF graphs and SPARQL to try to create my own knowledge engine.

Comment · 3 years ago

I live in DC, what meetups do you run? I would love to join