[CG] Latest from the world of LLMs

Michael D and Michael DB, The answer to that question is very clear and simple: MDB: Where is there a purely symbolic AI system that can support NLP anywhere in the same ballpark as LLMs? If such a thing exists would love to know about it. Keep in mind the system has to do a lot more than turn natural language into SPARQL. It needs to keep the context of a discussion thread, it needs to be able to respond appropriately when you give it feedback like: "give me more/less detail" "change the tone to [more playful] [more professional] [more creative]" etc. It needs to be able to do common sense reasoning and handle very hard NLP problems like anaphora. Answer: We did that with the purely symbolic VivoMind technology (from 2000 to 2010). The language understanding part was far better than the language generating part. In accuracy and precision, the 2010 version far exceeds any and every current LLM system that does not use a symbolic component to detect and evaluate the inevitable errors and hallucinations. Our largest clients (some US gov't agencies) attest to that fact. For an overview and documentation of the old VivoMind system, see https://jfsowa.com/talks/cogmem.pdf . The date of this presentation is 2016, but it is an update of slides that were originally presented in 2011. See the applications that begin on slide 45. Every one of them was implemented in 2010 or earlier. And in precision and accuracy, they are far ahead of anything that can be implemented by LLMs today without using symbolic technology for reasoning and evaluation. Our new Permion Inc. system combines an upgraded version of the 2010 symbolic technology with LLMs for an enhanced natural language OUTPUT that has a more readable and more elegant syntax. Permion also uses LLMs for some of the searching, but it only requires a modest amount of storage. For customers with deep pockets, however, the Permion technology can scale up to use as much storage as the customer provides. For a talk on YouTube, see Evaluating and Reasoning with and about GPT by John Sowa and Arun Majumdar, https://www.youtube.com/watch?v=6K6F_zsQ264 . That talk only had 544 views. But the following YouTube, from the 2024 Ontolog Summit, had over 10 thousand views: Without Ontology LLMs are Clueless by John Sowa, https://www.youtube.com/watch?v=t7wZbbISdyA . We strongly suspect that a large percentage of the viewers are in China. One reason is that I have been invited to give talks (expenses paid) in places like Hong Kong, Macau, Singapore... Another reason: Arun has also received an offer from China to buy our Permion company. However, some of our key customers would not be happy with that idea. John ---------------------------------------- From: "Michael Denny" <regattago(a)gmail.com> Michael, I was a little surprised to see you say “What LLMs do is to solve the problem of common sense reasoning better than any other approach to date. LLMs know that things fall down, that water is wet, etc. “ When I asked Claude 3.5 the following: “Is it accurate to say that LLMs solve the problem of common sense reasoning and that LLMs know that things fall down, that water is wet, etc.?” The response was: “No, that's not accurate. The relationship between LLMs and common sense reasoning is much more complex and limited. LLMs can often produce text that appears to reflect common sense understanding, but this is primarily because they've learned statistical patterns in human-written text about how objects and phenomena are typically described. They don't have genuine understanding or reliable common sense reasoning capabilities. For example, an LLM might correctly state that objects fall downward or that water makes things wet, but it doesn't truly "understand" gravity or wetness the way humans do. Some key limitations: - Inconsistency: LLMs can make basic physics or common sense errors when questions are asked in novel ways or contexts - No real-world grounding: Their "knowledge" comes from text patterns, not from physical experience with the world - No causal understanding: They don't truly grasp cause and effect relationships - Pattern matching vs. reasoning: They often give sensible-looking answers by matching patterns rather than through actual reasoning This is why LLMs can sometimes make surprisingly basic errors about physical reality or common sense relationships that would be obvious to humans. They might correctly state that "water flows downhill" in one context but then make an elementary mistake about fluid dynamics in another context. What are your thoughts on the nature of common sense reasoning in AI systems? I'm curious about your perspective on this.” I think the key point is that LLMs do not understand naïve physics or any deterministic relationships as humans do, nor have the inferential capabilities symbolic logics do. But as you point out, they do make a great interlocutor. Mike From: ontolog-forum(a)googlegroups.com [mailto:ontolog-forum@googlegroups.com] On Behalf Of Michael DeBellis Sent: Monday, February 3, 2025 6:47 PM To: ontolog-forum Subject: [ontolog-forum] Re: Latest from the world of LLMs ---------------------------------------- It is a big leap from what we know about DeepThink to make that conclusion. For one thing, I don't trust the mainstream press (CNBC) for any info about a complex scientific or technology question. For another, there is so much we don't know with confidence about DeepThink. How well does it really do on benchmarks? How much money or computing resources did the Chinese government or other groups associated with them pump into DeepThink? For another, there are things that Open AI can do that DeepThink can't do such as their recent Deep Research announcement: https://www.youtube.com/watch?v=YkCDVn3_wiw ; Also, see the Playlist "12 Days of Open AI" on Youtube: https://www.youtube.com/playlist?list=PLOXw6I10VTv9lin5AzsHAHCTrC7BdVdEM But even supposing that every claim made about DeepThink is true, it doesn't mean "the myth of an LLM Scaling Moat has been debunked" It just means the moat isn't as large as we thought. And if that moat isn't as large as we thought, I don't see how that justifies the conclusion: " All roads lead back to good old symbolic reasoning and inference—now more accessible and understandable when combined with a multi-modal natural language interface that enhances interaction possibilities." If anything, it makes LLMs even MORE usable and accessible because it doesn't require the huge resources that we used to think to create one (from what I know, I'm deeply skeptical but I actually hope that the claims of DeepThink are true because I think it is a good thing that we don't restrict foundational LLMs to just a few companies) Where is there a purely symbolic AI system that can support NLP anywhere in the same ballpark as LLMs? If such a thing exists would love to know about it. Keep in mind the system has to do a lot more than turn natural language into SPARQL. It needs to keep the context of a discussion thread, it needs to be able to respond appropriately when you give it feedback like: "give me more/less detail" "change the tone to [more playful] [more professional] [more creative]" etc. It needs to be able to do common sense reasoning and handle very hard NLP problems like anaphora. What LLMs do is to solve the problem of common sense reasoning better than any other approach to date. LLMs know that things fall down, that water is wet, etc. It doesn't make sense to also expect them to be able to solve that problem AND have completely reliable domain knowledge. There are several approaches to using curated, authoritative knowledge sources with an LLM to reduce or eliminate hallucinations. One of them is Retrieval Augmented Generation (RAG). That's an example of how semantic tech and LLMs complement each other: https://journals.sagepub.com/eprint/RAQRDSYGHKBSF4MRHZET/full also see: https://www.michaeldebellis.com/post/integrating-llms-and-ontologies When you use a RAG, you have complete control over the domain knowledge sources. Building a RAG is the best way I've ever seen to put an NLP front end to your knowledge graph. Another way to get domain knowledge into an LLM are to train LLMs with special training sets for the domain. One excellent example of this is Med-BERT, a version of the BERT LLM specially trained with healthcare data. See: L. Rasmy, Y. Xiang, Z. Xie, C. Tao and D. Zhi, "Med-BERT: pretrained contextualized embeddings on large-scale structured electronic health records for disease prediction," Nature npj digital medicine, vol. 4, no. 86, 2021. Still another (much newer) way is to use reinforcement fine-tuning to train an LLM with domain specific data. This doesn't take as much work as the domain specific LLM but is much more work than creating a RAG. See: M. Chen, J. Allard, J. Wang and J. Reese, "Reinforcement Fine-Tuning—12 Days of OpenAI: Day 2," 6 December 2024. https://www.youtube.com/watch?v=yCIYS9fx56U&t=29s