Andrew Ng

[Michael DeBellis]

In the meeting today I mentioned Andrew Ng as an example of an ML researcher at Stanford that I think it might be interesting to reach out to. Ravi asked me to post some links to Andrew's work. He has a newsletter that I read every week and is filled with information that is both excellent research and infused with an understanding of business realities. I first came across Andrew when I was looking to understand ML and took his excellent course on the topic on Coursera several years ago. 

Here is the home page for his group: https://www.deeplearning.ai/the-batch/ You can sign up for his newsletter at the bottom of the page. 

Recently he posted a series of articles on LLMs and Intelligent Agents. This is a topic I think is fascinating and also another area where I think there is a lot of potential for systems that utilize LLM and knowledge graph technology. He lists the four essential topics for LLM powered agents as: 1) Reflection 2) Tool use 3) Planning and 4) Collaborating distributed agents. 

Here is the first post on Agents, there are links here to the others and also several papers on each topic that go into much more depth: 

https://www.deeplearning.ai/the-batch/how-agents-can-improve-llm-performance/?utm_campaign=The%20Batch&utm_medium=email&_hsenc=p2ANqtz-9XPEoXqKMrZrh0TamVPYc7dY05ZXTLyhuXdQarX52GWJl2G1X2cked2gc-i_GKAWK4JIsb11FpDZbntRm-keFAtJlyQw&_hsmi=305213422&utm_content=305213422&utm_source=hs_email 

[Ravi Sharma]

Micheal

Yes A couple of speakers did talk about specific roles and use of Agents, but including Fall series on LLMs and Ontologies and current Summit 2024, their mention has been not too often.

What message do you think ANdrew Ng would give to our summit in a paragraph?

I will try to look at the link as well.

Thanks. Regards.

Ravi

(Dr. Ravi Sharma, Ph.D. USA)

NASA Apollo Achievement Award

Former Scientific Secretary iSRO HQ

Ontolog Board of Trustees

Particle and Space Physics

Senior Enterprise Architect

SAE Fuel Cell Standards Member

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/0f17d6fc-1c85-48ad-b784-3677b296d77cn%40googlegroups.com.

[Michael DeBellis]

What message do you think ANdrew Ng would give to our summit in a paragraph?

Ravi, I honestly don't know. There have been a couple presentations where I thought that semantic technology was highly relevant to what he was talking about. One of them was on improving ML systems by better training data. He spent a lot of time talking about the point that the best way to increase the accuracy of an ML system was not just increasing the amount of the data but increasing the quality of the data. That was a topic where I thought Semantic technology could be especially relevant in many ways. Defining data integrity constraints with SHACL and throwing out data that violates those constraints as well as using OWL models to better understand the data and using the ontology to evaluate specific elements (i.e., instances in a knowledge graph) to again determine if certain examples are probably the result of errors. Also, do identify trends in the data that might help provide better structure to the ML models. I've thought about reaching out to professor Ng. In the past I've done that with a few well known academics like Chomsky and I usually get very favorable responses. Several years ago I had some amazing detailed discussions with Chomsky about issues related to ethics, linguistics, and theory of mind. But Ng is very much in demand by various businesses as well as being a respected academic and my feeling is he probably gets swamped with so many random emails from undergrads learning ML that he's less likely to reply. 

The main point I was trying to make in the meeting is IMO right now there is a large disparity between the ML people and the Semantic side and my impression is the ML people barely know we still exist. I don't know how to address that but I think it is worth thinking about some potential joint work with ML leaders like Ng and people from Ontolog in the future. 

Michael

You received this message because you are subscribed to a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/457W-hZDHaU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5fD%3DXBAxQP2jCXUg--Oob1POMooJ6fKLusbnnFbAcyZ2g%40mail.gmail.com.

[Ravi Sharma]

Michael

Great suggestions and we will try to invite him in our future summit sessions or as a special speaker as we have done with John, Berry and others.

As Ken Gary and I have this email, we will use a couple of sentences relevant to communique.

I have a professional clarification, I think of myself as a student of this amazing concept of Ontology which most in the west have learned from Aristotle and Greeks. I view similar knowledge based on my background that touches Sanskrit that has similar but different processes and in the very long term I intend to contribute to bring out similarities and differences, that certainly include discussions between integral and differential points of view, there are many others.

I understand a bit of ML, but not necessarily belong to the semantic community. Slightly different meanings of a term such as semantics does not take ontologists away from Knowledge related aspects.

Appreciate your inputs.

Thanks.

Ravi

(Dr. Ravi Sharma, Ph.D. USA)

NASA Apollo Achievement Award

Former Scientific Secretary iSRO HQ

Ontolog Board of Trustees

Particle and Space Physics

Senior Enterprise Architect

SAE Fuel Cell Standards Member

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFikey6ZRmXR9-Fu5R1B1HLP2umAhbUfNELnyyfq9kfs7Hfg%40mail.gmail.com.

[Michael DeBellis]

 I view similar knowledge based on my background that touches Sanskrit that has similar but different processes

Ravi, You may have seen this but I remember back when I used to read AI magazine from cover to cover there was a fascinating article on Sanskrit and Knowledge Representation. I don't remember the details but I found it online if anyone is interested: https://ojs.aaai.org/aimagazine/index.php/aimagazine/article/view/466  

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAAN3-5fJVU0%2B52-gAg0K-ZXVUzOiNDVoqf9Dn0sHm%2B4VX8NGxg%40mail.gmail.com.

[John F Sowa]

Ravi,  Michael, List,

Ravi:  I think of myself as a student of this amazing concept of Ontology which most in the west have learned from Aristotle and Greeks. 

So do I.  In fact, Aristotle and the Stoics specified an excellent replacement for OWL.  Aristotle organized his categories in a hierarchy that can be represented by the hierarchy part of OWL (the best and most useful part of OWL).  The four sentence types for his logic (A, I, E, O) specify the operations for reasoning with and about the hierarchy.

Then the Stoics specified if-then rules for stating constraints and performing inferences with and about the hierarchy and related information.  Although they didn't have full FOL, their if-then rules were later extended to full FOL by the medieval logicians.  Ockham had specified full FOL in the 14th century, but stated in a version of Latin.  

That combination would have been equivalent to KL-ONE (by Brachman and Fikes in 1973).  They coined the term T-Box (Terminology box) for the hierarchy and A-Box (Assertion box) for the equivalent of FOL.  Since they did not restrict their specifications by decidability, their T-Box plus A-Box would be equivalent to  "good OWL" -- i.e. OWL without the limitations of decidability.   Unfortunately, Brachman later listened to the decidability gang who inflicted decidabilit .  That was a major step backward from the middle ages.

The following point is fundamental to knowledge representation of any kind, especially applications of ontology:

Michael DB:  The main point I was trying to make in the meeting is IMO right now there is a large disparity between the ML people and the Semantic side and my impression is the ML people barely know we still exist. I don't know how to address that but I think it is worth thinking about some potential joint work with ML leaders like Ng and people from Ontolog in the future. 

One of the great strengths of Ontolog Forum is an issue that many (most?) theoreticians ignore:   How do you use logic to do anything useful?   Many excellent theoreticians don't have any experience in practical applications.

I admit that was my outlook when I was studying math & logic.  But at IBM, my first job was in a mathematical analysis group, where we had to solve problems for which the engineers didn't have the training.  So I had to learn a great deal about their problems and about practical computational methods for solving them.  That was excellent on-the-job training for which I had to educate myself.

Michael:  I think it is worth thinking about some potential joint work with ML leaders like Ng and people from Ontolog in the future. 

But it's important to recognize the goals and interests of any people you invite.  You may have learned a great deal about linguistics by talking with Chomsky.  But it's unlikely that Chomsky would take much interest in studying any problems you may have.  (I know quite  a few people who have been badly bruised or burned by Chomsky because they dared to modify his theories in order to deal with practical problems.)

Re Andrew Ng:  I have read some of his writings, and I agree that they're good.  But as you said in your note, "the ML people barely know we [ontologists] still exist. I don't know how to address that."  

For some background on the history of AI, I recommend an article from 1983:  Krypton: A Functional Approach to Knowledge Representation,  https://web.stanford.edu/class/cs227/Readings/KryptonAFunctionalApproachToKRR.pdf

They discuss the issues and some of the historical developments.  Looking backwards from today, I would summarize the issues in three parts: 

1. An ontology specified as a hierarchy (tree or lattice).  This would correspond to Aristotle's hierarchy with his four rules for specifying and reasoning with and about it.  It would also correspond to the T-box.

2. Constraints on the hierarchy specified in full FOL instead of the subset in OWL specified by Turtle or other notations.

3. Assertions in a database of SQL, object-oriented DB, or RDF assertions anywhere on the WWW.

Then add the constraint that #1 has priority over #2 in case of any contradictions, and #2 would have priority over assertions in #3 that come from a DB (SQL or object-oriented).    For assertions that come from anywhere else, there may be good reasons for giving them higher priority, but that would depend on the sources and the circumstances.

John

 

---

From: "Michael DeBellis" <mdebe...@gmail.com>

[Marcelino Borges]

Furthermore, although Andrew Ng works mainly with machine learning, I saw a recent talk of his in which he comments that AI is much more than that, emphasizing the role of explicit knowledge models in impacting many businesses.

--

[Ravi Sharma]

Marcelino

Yes I always thought so.

Thanks.

Ravi

(Dr. Ravi Sharma, Ph.D. USA)

NASA Apollo Achievement Award

Former Scientific Secretary iSRO HQ

Ontolog Board of Trustees

Particle and Space Physics

Senior Enterprise Architect

SAE Fuel Cell Standards Member

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CANpcTFq03Qxx04oE%3DkDRRExDUaDGh1FVDuTZ%3D6GMfzk%3DqUS_0Q%40mail.gmail.com.

[Ravi Sharma]

John - I have a further question.

I replied only to our group.

Regarding Point 3, 

• I think it includes all types of relationships, not only hierarchical ones!
• For random entities, we might use Queries and LLMs to reason and discover potentially valid relationships?
• How are cases dealt with in ontologies if not as relationship types that need not be hierarchical?

Regards.

Thanks.

Ravi

(Dr. Ravi Sharma, Ph.D. USA)

NASA Apollo Achievement Award

Former Scientific Secretary iSRO HQ

Ontolog Board of Trustees

Particle and Space Physics

Senior Enterprise Architect

SAE Fuel Cell Standards Member

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/e69f3c928be04667ac66eec6080bee93%40bestweb.net.