ConferenceCall 2025 02 26: Difference between revisions
Ontolog Forum
(→Agenda) |
|||
| (6 intermediate revisions by the same user not shown) | |||
| Line 23: | Line 23: | ||
* '''[[ArunMajumdar|Arun Majumdar]]''' and '''[[JohnSowa|John Sowa]]''' - ''Generating Ontologies'' | * '''[[ArunMajumdar|Arun Majumdar]]''' and '''[[JohnSowa|John Sowa]]''' - ''Generating Ontologies'' | ||
* Abstract: The LLMs of Generative AI support powerful methods for translating, finding, combining, and transforming data represented in languages of any kind — natural, artificial, linear, diagrammatic, pictorial, or structural in any number of dimensions. Their reasoning is based on abduction or educated guessing. But abduction alone cannot evaluate the accuracy of the guesses. Therefore, many generative AI systems are hybrids that also use traditional symbolic AI methods for detecting and avoiding errors. Permion technology is based on an integrated combination of the best generative and symbolic methods. Reasoning is performed in a cognitive cycle of abduction, deduction, testing, induction, and repeat. The cognitive cycle begins with a basic ontology, which it revises and extends dynamically. | * Abstract: The LLMs of Generative AI support powerful methods for translating, finding, combining, and transforming data represented in languages of any kind — natural, artificial, linear, diagrammatic, pictorial, or structural in any number of dimensions. Their reasoning is based on abduction or educated guessing. But abduction alone cannot evaluate the accuracy of the guesses. Therefore, many generative AI systems are hybrids that also use traditional symbolic AI methods for detecting and avoiding errors. Permion technology is based on an integrated combination of the best generative and symbolic methods. Reasoning is performed in a cognitive cycle of abduction, deduction, testing, induction, and repeat. The cognitive cycle begins with a basic ontology, which it revises and extends dynamically. | ||
* [https://ontologforum.s3.us-east-1.amazonaws.com/OntologySummit2025/Track-2/GeneratingOntologies--JohnSowa-ArunMajumdar_20250226.pdf Slides] | |||
* [https://ontologforum.s3.us-east-1.amazonaws.com/OntologySummit2025/Track-2/GeneratingOntologies--JohnSowa-ArunMajumdar_20250226.mp4 Video Recording] | |||
* [https://youtu.be/zRYJE6QJZx0 YouTube Video] | |||
== Conference Call Information == | == Conference Call Information == | ||
| Line 31: | Line 34: | ||
{{:OntologySummit2025/ConferenceCallInformation}} | {{:OntologySummit2025/ConferenceCallInformation}} | ||
== | == Discussion == | ||
12:13 Ravi Sharma: John, dynamism is all around us hence descriptive Ontologies require dynamic modeling, how do graphs also changing with dynamism represent descriptive ontology, except as akin to videos? | |||
* 12:29 Alican Tüzün: 👍 | |||
12:18 Alican Tüzün: Is the cognitive Memory, like a set of facts from the world? | |||
12:27 Douglas Miles: /me loud applause | |||
12:28 Ravi Sharma: Semantics of Language as you describe it is Important – just between OODA and Pierces’ 4 adjectives is non-trivial? Thus Languages and Meaning are non-unique transformations.. | |||
* 12:29 Alican Tüzün: 👍 | |||
12:29 Jack Park: Minsky’s book “The emotion machine” | |||
12:30 Ravi Sharma: The Verb Origin (root) is connected to intended meaning. | |||
12:34 John Sowa: The most important aspect of cognitive memory is the integratedmethod for using information of any kind from every source. | |||
* 12:36 Alican Tüzün: why not select the "useful" information which is something to do with the question which you designed your system for? | |||
12:34 Alican Tüzü: The computational logical language i presume FOL? | |||
12:37 John Sowa: Permion uses a superset of IS standard Prolog. by defaut it supports FOL. But it also supports metalanguage, which supports a huge amount of extensions, both precise and scruffy. | |||
* 12:38 Alican Tüzün: 👍 | |||
12:40 Alican Tüzün: because the native FOL does not have temporality, set-data relationships etc? I am asking because I am also building something similar myself for my PhD work, and came to similar limitations | |||
12:41 ToddSchneider: What is the extent of a ‘seed ontology’? | |||
* 12:42 Alican Tüzün: If it's same as what I am doing, its just V0 theory about the phenomena which is being or wanted to be observed | |||
12:45 John Sowa: Metalanguage can support anything - -- modal, temporal, probability scruffy, etc. Send a question to Ontolog forum about that.- -- | |||
* 12:47 Alican Tüzün: 👍 | |||
12:48 Ravi Sharma: Arun and John - Without your architecture you could not have untangled deepseek related understanding of how they probably have architected Deepseek? | |||
12:49 Gary Berg-Cross: As a baseline does application of Permion using your own documents give you what you would call and accurate view of the "architecture"? | |||
12:49 Paul Brown (Permion): Great talk, thanks John & Arun. Nice user-interface 😉 | |||
12:55 Douglas Miles: if the Arity of the types of Queries get large due to being mapped from Prolog | |||
12:55 ToddSchneider: My question to John and Arun was “What is the extent of that ontology and If you're using first order logic for it or prologue, is there a bound on the arity of the relations?" | |||
13:03 Phil Jackson: Why is deepseek apparently getting better performance than its competitor systems? | |||
13:10 Mike Peters: It was a fascinating presentation. Thank you Arun and John | |||
* 13:10 Jack Park: 💯 | |||
* 13:16 Alican Tüzün: Thank you for the insight | |||
* 13:16 Mike Peters: 👍 | |||
13:13 Mark Underwood: Future versions should interface to Rust | |||
13:13 Phil Jackson: I was just asking a sort-of bottom-line question about deepseek - thanks for a very interesting presentation | |||
13:16 Mark Underwood: Thanks, all | |||
13:16 Jack Park: https://en.wikipedia.org/wiki/X-machine | |||
== Resources == | == Resources == | ||
* [https://ontologforum.s3.us-east-1.amazonaws.com/OntologySummit2025/Track-2/GeneratingOntologies--JohnSowa-ArunMajumdar_20250226.pdf Slides] | |||
* [https://ontologforum.s3.us-east-1.amazonaws.com/OntologySummit2025/Track-2/GeneratingOntologies--JohnSowa-ArunMajumdar_20250226.mp4 Video Recording] | |||
* [https://youtu.be/zRYJE6QJZx0 YouTube Video] | |||
== Previous Meetings == | == Previous Meetings == | ||
Latest revision as of 00:57, 8 April 2025
| Session | Track 2 |
|---|---|
| Duration | 1 hour |
| Date/Time | 26 Feb 2025 17:00 GMT |
| 9:00am PST/12:00pm EST | |
| 5:00pm GMT/6:00pm CET | |
| Convener | Alex Shkotin |
Ontology Summit 2025 Track 2
Agenda
- Arun Majumdar and John Sowa - Generating Ontologies
- Abstract: The LLMs of Generative AI support powerful methods for translating, finding, combining, and transforming data represented in languages of any kind — natural, artificial, linear, diagrammatic, pictorial, or structural in any number of dimensions. Their reasoning is based on abduction or educated guessing. But abduction alone cannot evaluate the accuracy of the guesses. Therefore, many generative AI systems are hybrids that also use traditional symbolic AI methods for detecting and avoiding errors. Permion technology is based on an integrated combination of the best generative and symbolic methods. Reasoning is performed in a cognitive cycle of abduction, deduction, testing, induction, and repeat. The cognitive cycle begins with a basic ontology, which it revises and extends dynamically.
- Slides
- Video Recording
- YouTube Video
Conference Call Information
- Date: Wednesday, 26 February 2025
- Start Time: 9:00am PST / 12:00pm EST / 6:00pm CET / 5:00pm GMT / 1700 UTC
- ref: World Clock
- Expected Call Duration: 1 hour
- Video Conference URL: https://us02web.zoom.us/j/88593616861?pwd=HafnK0yB7PFDK1EyiUyQRDKanZlbjU.1
- Conference ID: 885 9361 6861
- Passcode: 306236
Discussion
12:13 Ravi Sharma: John, dynamism is all around us hence descriptive Ontologies require dynamic modeling, how do graphs also changing with dynamism represent descriptive ontology, except as akin to videos?
- 12:29 Alican Tüzün: 👍
12:18 Alican Tüzün: Is the cognitive Memory, like a set of facts from the world?
12:27 Douglas Miles: /me loud applause
12:28 Ravi Sharma: Semantics of Language as you describe it is Important – just between OODA and Pierces’ 4 adjectives is non-trivial? Thus Languages and Meaning are non-unique transformations..
- 12:29 Alican Tüzün: 👍
12:29 Jack Park: Minsky’s book “The emotion machine”
12:30 Ravi Sharma: The Verb Origin (root) is connected to intended meaning.
12:34 John Sowa: The most important aspect of cognitive memory is the integratedmethod for using information of any kind from every source.
- 12:36 Alican Tüzün: why not select the "useful" information which is something to do with the question which you designed your system for?
12:34 Alican Tüzü: The computational logical language i presume FOL?
12:37 John Sowa: Permion uses a superset of IS standard Prolog. by defaut it supports FOL. But it also supports metalanguage, which supports a huge amount of extensions, both precise and scruffy.
- 12:38 Alican Tüzün: 👍
12:40 Alican Tüzün: because the native FOL does not have temporality, set-data relationships etc? I am asking because I am also building something similar myself for my PhD work, and came to similar limitations
12:41 ToddSchneider: What is the extent of a ‘seed ontology’?
- 12:42 Alican Tüzün: If it's same as what I am doing, its just V0 theory about the phenomena which is being or wanted to be observed
12:45 John Sowa: Metalanguage can support anything - -- modal, temporal, probability scruffy, etc. Send a question to Ontolog forum about that.- --
- 12:47 Alican Tüzün: 👍
12:48 Ravi Sharma: Arun and John - Without your architecture you could not have untangled deepseek related understanding of how they probably have architected Deepseek?
12:49 Gary Berg-Cross: As a baseline does application of Permion using your own documents give you what you would call and accurate view of the "architecture"?
12:49 Paul Brown (Permion): Great talk, thanks John & Arun. Nice user-interface 😉
12:55 Douglas Miles: if the Arity of the types of Queries get large due to being mapped from Prolog
12:55 ToddSchneider: My question to John and Arun was “What is the extent of that ontology and If you're using first order logic for it or prologue, is there a bound on the arity of the relations?"
13:03 Phil Jackson: Why is deepseek apparently getting better performance than its competitor systems?
13:10 Mike Peters: It was a fascinating presentation. Thank you Arun and John
- 13:10 Jack Park: 💯
- 13:16 Alican Tüzün: Thank you for the insight
- 13:16 Mike Peters: 👍
13:13 Mark Underwood: Future versions should interface to Rust
13:13 Phil Jackson: I was just asking a sort-of bottom-line question about deepseek - thanks for a very interesting presentation
13:16 Mark Underwood: Thanks, all
13:16 Jack Park: https://en.wikipedia.org/wiki/X-machine
Resources
Previous Meetings
| Session | |
|---|---|
| ConferenceCall 2025 02 19 | Track 1 |
| ConferenceCall 2025 02 12 | Track 1 |
| ConferenceCall 2025 02 05 | Track 1 |
| ... further results | |
Next Meetings
| Session | |
|---|---|
| ConferenceCall 2025 03 05 | Track 2 |
| ConferenceCall 2025 03 12 | Track 2 |
| ConferenceCall 2025 03 19 | Synthesis |
| ... further results | |