Rob, basically you're reiterating what I've been saying here all along. To increase contextualization and instill robustness in the LLM systemic hierarchies. Further, that it seems to be critically lacking within current approaches.
However, I think this is fast changing, and soon enough, I expect breakthroughs in this regard. Neural linking could be one of those solutions. While it may not be exactly the same as your hypothesis (?), is it because it's part of your PhD that you're not willing to acknowledge that this theoretical work may have been completed by another researcher more than 17 years ago, even submitted for review and subsequently approved? The market, especially Japan, grabbed this research as fast as they could. It's the West that turned out to be all "snooty" about its meaningfulness, yet, it was the West that reviewed and approved of it. Instead of serious collaboration, is research not perhaps being hamstrung by the NIH (Not Invented Here) syndrome, acting like a stuck handbrake? IMO, this is exactly why progress in the West has been so damned slow. Everyone is competing for the honor of discovering something great, looking out for number one. Write a book. Grab a TV show, become a hero, or whatever. Here's another person on this list who also jetted off into his own space with a PhD, the fruits of his labor we have never seen. His theory becoming quickly as irrelevant as the time it takes to code newer applications. As you must be acutely aware of: Valid doesn't equate to Reliable and Reliable doesn't equate to Relevance (Attention) and Relevance doesn't equate to Intel. Seems to me, pundits of LLMs (used to) think it does. After relevance within LLMs would be resolved (soon), the real battle for Intel would begin. An epic battle well-worth watching. Meanwhile, in all probability, whatever we could think of/invent has already been thought of by another person, somewhere else in the world. Sometimes, centuries ago. This seems very similar to what Karl Mannheim was referring to in his view on competition for knowledge as a factor for human survival, within the context of the sociology of knowledge. I think we should add this as a mitigating factor to culturally-based knowledge systems and embed it in LLMs. My 2 cents' worth. Predictably, at a certain "size" LLMs - on their own - would wander off into ambiguity. There are multiple reasons for this, one of them due to exponential complexity. That's the point - I'll predict - at which ~99.7% of the LLM-dev market's going to be left behind to scramble for marketing-related contracts/jobs in order to support AI-based sales and trading efforts. The remainder ~0.3% are going to emerge as the AI-driven Intel industry. I'll rate the AI-Intel market segment as a future, trillion-dollar industry. Just some thoughts I had. I could be completely wrong. Only time would tell. On Sat, Jun 15, 2024 at 9:42 AM Rob Freeman <chaotic.langu...@gmail.com> wrote: > On Sat, Jun 15, 2024 at 1:29 AM twenkid <twen...@gmail.com> wrote: > > > > ... > > 2. Yes, the tokenization in current LLMs is usually "wrong", ... it > should be on concepts and world models: ... it should predict the > *physical* future of the virtual worlds > > Thanks for comments. I can see you've done a lot of thinking, and see > similarities in many places, not least Jeff Hawkins, HTM, and > Friston's Active Inference. > > But I read what you are suggesting as a solution to the current > "token" problem for LLMs, like that of a lot of people currently, > LeCun prominently, to be that we need to ground representation more > deeply in the real world. > > I find this immediate retreat to other sources of data kind of funny, > actually. It's like... studying the language problem has worked really > well, so the solution to move forward is to stop studying the language > problem! > > We completely ignore why studying the language problem has caused such > an advance. And blindly, immediately throw away our success and look > elsewhere. > > I say look more closely at the language problem. Understand why it has > caused such an advance before you look elsewhere. > > I think the reason language models have led us to such an advance is > that the patterns language prompts us to learn are inherently better. > "Embeddings", gap fillers, substitution groupings, are just closer to > the way the brain works. And language has led us to them. > > So OK, if "embeddings" have been the advance, replacing both fixed > labeled objects in supervised learning, and fixed objects based on > internal similarities in "unsupervised" learning, instead leading us > to open ended categories based on external relations, why do we still > have problems? Why can't we structure better than "tokens"? Why does > it seem like they've led us the other way, to no structure at all? > > My thesis is actually pretty simple. It is that these open ended > categories of "embeddings" are good, but they contradict. These "open" > categories can have a whole new level of "open". They can change all > the time. That's why it seems like they've led us to no structure at > all. Actually we can have structure. It is just we have to generate it > in real time, not try to learn it all at once. > > That's really all I'm saying, and my solution to the "token" problem. > It means you can start with "letter" tokens, and build "word" tokens, > and also "phrases", whole hierarchies. But you have to do it in real > time, because the "tokens", "words", "something", "anything", "any > thing", two "words", one "word"... whatever, can contradict and have > to be found always only in their relevant context. > > Do you have any comments on that idea, that patterns of meaning which > can be learned contradict, and so have to be generated in real time? > > I still basically see nobody addressing it in the machine learning > community. > > It's a little like Matt's "modeling both words and letters" comment. > But it gets beneath both. It doesn't only use letters and words, it > creates both "letters" and "words" as "fuzzy", or contradictory, > constructs in themselves. And then goes on to create higher level > structures, hierarchies, phrases, sentences, as higher "tokens", > facilitating logic, symbolism, and all those other artifacts of higher > structure which are currently eluding LLMs. All levels of structure > become accessible if we just accept they may contradict, and so have > to be generated in context, at run time. > > It's also not unrelated to James' definition of "a thing as anything > that can be distinguished from something else." Though that is more at > the level of equating definition with relationship, or "embedding", > and doesn't get into the missing, contradictory, or "fuzzy" aspect. > Though it allows that fuzzy aspect to exist, and leads to it, if once > you imagine it might, because it decouples the definition of a thing, > from any single internal structure of the thing itself. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T682a307a763c1ced-M822a6c5e20cedd6f6974bf19 Delivery options: https://agi.topicbox.com/groups/agi/subscription
