Connectionists: Can LLMs think?

[Ali Minai]

Paul

Thank you for what I think is one of the most insightful critiques of LLMs
that I have seen on this - or any - thread. We keep forgetting that
intelligence is fundamentally a *biological* phenomenon, not a
computational one, though, of course, we can choose to see biological
phenomena in the framework of computation. The function of intelligence and
all its components - pattern recognition, understanding, inference,
reasoning - is ultimately grounded in its contribution to the survival of
the animal, and that is - as you argue - the source of meaning. By turning
linguistic expression into a symbol generating computational process rather
than a purposive biological process, what we get is, at best, a very
convincing simulacrum of intelligent expression grounded in text (and
code), not reality. And, like all simulacra, it breaks down absurdly when
it reaches the limits of its shallow grounding. In contrast, when animal
intelligence breaks down - and it does - it is in ways consistent with
reality and, therefore, understandable. After all, the animal is not
expressing itself based on a superficial representation created from a data
corpus of one or two modalities, but from its completely multimodal,
continual, and always-integrated experience of reality. That is the reason
why causality is built into its intelligence - it reflects the causality of
the real world. That is also why is can move "out of sample" much more
gracefully because, barring black swan events, even novel experiences are
seamlessly connected with its prior experiences.

A key point in all this is the fact that the animal's intelligence is not
based on priors configured just from a dataset, but those configured by a
deep hierarchy of processes: Evolution, development, and learning, leading
to real-time adaptive emergent behaviors some of which in complex animals
such as humans are immensely creative. It's the
perceptual-cognitive-behavioral capacity shaped by this deep adaptive
hierarchy that enables animals to produce complex behavior (including
linguistic expression) that we se as intelligent. In contrast, we wish to
train initially naive systems such as randomly initialized neural networks
to acquire this capacity purely by being trained on data that is itself not
reality but only a derivative of it. The fact that something like ChatGPT
can achieve what it has in this situation is amazing, and, I believe, does
tell us something about human language. But it is not even close to forming
the kernel of a true intelligence in the natural sense. Now, if we were to
embed something like a multimodal and much larger version of ChatGPT into a
carefully designed robot (to mimic the evolutionary aspect), and allow it
to learn in stages like animals do (to mimic the developmental part),
things could get very interesting. We probably would not want such robots
to do their learning out in the real world, so perhaps very strong VR would
have to be part of this scenario. All this, of course, only if we want to
produce such autonomous beings and loose them upon the world. One thing is
certain, though: If we do end up building such systems, we should not
expect them to be perfectly obedient, predictable, or transparent. Those
things are just not compatible with real intelligence.

Ali



*Ali A. Minai, Ph.D.*
Professor and Graduate Program Director
Complex Adaptive Systems Lab
Department of Electrical & Computer Engineering
828 Rhodes Hall
University of Cincinnati
Cincinnati, OH 45221-0030

Phone: (513) 556-4783
Fax: (513) 556-7326
Email: Ali.Minai at uc.edu
          minaiaa at gmail.com

WWW: https://eecs.ceas.uc.edu/~aminai/ <http://www.ece.uc.edu/%7Eaminai/>


On Fri, Mar 24, 2023 at 11:36 AM Paul Cisek <paul.cisek at umontreal.ca> wrote:

> I must admit that it’s quite intimidating to be disagreeing with Geoff
> Hinton, whom I’ve admired for so many years… But it’s just a difference of
> opinion, so here goes:
>
>
>
> I was not suggesting there is something special about people. In fact,
> it’s rather the opposite: I think that the similarity between human and
> animal brains/behaviors strongly suggests that whatever our kind of
> intelligence involves is actually based on fundamental processes that we
> share with many animals, but not with LLMs. One critical issue is related
> to the question of meaning.
>
>
>
> The brain is not a general information processing system whose task is to
> produce an output given an input. It’s a more specific thing: a control
> system that controls its input via output through the environment. It
> establishes adaptive closed-loop interaction with a complex world by
> complementing the dynamics of that world in such a way that the resulting
> animal-environment system tends toward states that are desirable for the
> animal (i.e., being well fed, not under threat, etc.). This could involve
> all kinds of internal variables that covary with the external world, in
> some cases simply because they are dynamically coupled with the world, and
> in some cases because they capture generalizable information about that
> world (i.e. what we might call “knowledge”). Obviously, adaptive behavior
> is the whole point of having a brain, so I doubt anyone would really argue
> with this point… except perhaps to say that it is trivial and you’ve heard
> it all before.
>
>
>
> Sure, it’s trivial and obvious, but it has some important implications
> worth repeating. In particular, certain interactions are meaningful for the
> organism in that they positively or negatively influence its state.
> Ingesting food reduces hunger, running away from a predator reduces danger,
> etc. You could say we can adaptively control our state because of the
> predictable consistency of the world in which we live, which we can use to
> our advantage. For this reason certain actions we perform, and certain
> sensory information we use to select among and to guide those actions, have
> meaning to us. Well it’s the same for communication. When in the presence
> of complex creatures, we can control our state via sounds or gestures,
> because those sounds or gestures can reliably persuade those complex
> creatures to act in a way that benefits us. Consider a human infant that
> cannot accomplish much on its own. Fortunately, it so happens that in the
> niche of helpless human infants there is something called a parent, which
> has the handy properties of being incredibly complex but also very easy for
> the infant to control. The baby cries and the parent rushes over to figure
> out and fix whatever is the problem, whether this involves getting some
> milk or driving at high speed to a hospital. With time, the baby can learn
> to make different noises to produce different outcomes via the parent, and
> in fact the parent will deliberately help the baby learn which noises make
> the parent bring it food versus water versus changing the diaper, etc.
> Throughout, the real purpose of making the noises is not to convey
> knowledge but to persuade. Animals do this all the time, from the threat
> postures of crayfish to monkeys baring their teeth to humans uttering “back
> off!”. The important point here is that the meaning is not in the syntax of
> a given utterance, but in the interaction that the utterance is likely to
> induce in those who speak the same language, and the desired consequence of
> that interaction. The words are just a kind of “shorthand notation” for the
> meaningful interactions, and they are compact and “symbolic” because the
> complex external agent will handle the details of implementing the
> interaction. Human language takes this to extremes of abstraction, but the
> fundamental context of control is still there. Even this type of
> “philosophical” discussion is an example of my attempt at persuasion.
>
>
>
> So is there anything like that in LLMs? To my understanding, no, because
> what LLMs learn to produce are just the words themselves without connection
> to the meaningful interactions for which those words are mere shorthand
> notation. The data on which LLMs are trained is text that is meaningful to
> the humans who produced it, and for that reason the text has a particular
> structure that follows from the meaningful interactions the humans were
> talking about (e.g. causes precede effects, paragraphs stay on topic,
> certain things get repeated, etc.). With enough training, the LLM can
> capture that structure by factoring out the interactions among the discrete
> symbols and clusters of symbols, and produce utterances that mimic it by
> applying those same kinds of patterns at multiple scales (phrases,
> sentences, paragraphs, and even writing style). But the actual semantic
> content is always outside of the system. The human who reads the text
> produced by the LLM will be able to guess at the original meaning that the
> words refer to (and will be predisposed to ascribe intention and purpose to
> the LLM), but the system itself has no access to that. In short: The
> meaning to which the words refer is understood by the humans who produced
> the training text, and by the humans who read the transformed text, but the
> transformer of the text itself never had any connection to the interactions
> that gave the words their meaning. That’s why it fails so often in
> conditions that we find trivial.
>
>
>
> I suppose there are many way to define what “thinking” is, or what
> “understanding” is, and perhaps there are general and useful definitions
> based on syntactic and structural richness that LLMs would satisfy. But if
> the question is whether they think in the same way as we do, then I believe
> the answer is no. Admittedly, we don’t know how our thinking is implemented
> by our brains. But even if you don’t have complete knowledge of something,
> you can still reject hypotheses that are incompatible with the knowledge
> that you do have. I believe we know enough about the brain and behavior (of
> humans and animals) that we can be confident that however our “thinking” is
> implemented, it is not based solely on the linguistic noises we utter but
> on how we learned to use those noises to construct complex interactions
> among ourselves.
>
>
>
> Finally, let me clarify that I’m not trying to shoot arrows into anyone’s
> back or to discourage efforts for building “strong AI”. It’s the opposite:
> What I’m trying to do is to *persuade* people to not neglect the issue of
> meaning because I don’t believe we can build humanlike intelligence without
> it. Focusing purely on syntax has led to dead ends many times already.
> There are other directions that I think are more promising, for example, in
> theories of closed-loop control systems that discover adaptive interactions
> using action-perception cycles and reinforcement learning. Many of these
> perhaps also have excess reliance on gigantic training sets, but at least
> the framing of the problem is one that has a chance of capturing the
> complex meaningful interactions first, and then only later seek to
> construct abstracted hierarchies of control and short-hand notations for
> those meaningful interactions. And no, I don’t think you need to be
> biological or “embodied” to have intelligence, but you do need to keep
> meaning within the system.
>
>
>
> Paul Cisek
>
>
>
>
>
>
>
> *From:* Geoffrey Hinton <geoffrey.hinton at gmail.com>
> *Sent:* Monday, March 20, 2023 1:59 PM
> *To:* Paul Cisek <paul.cisek at umontreal.ca>
> *Cc:* connectionists at mailman.srv.cs.cmu.edu
> *Subject:* Re: Connectionists: Can LLMs think?
>
>
>
> LLM's do not do pattern matching in the sense that most people understand
> it. They use the data to create huge numbers of features and
> interactions between features such that these interactions can predict the
> next word.
>
> The first neural net language model (so far as I know) made bets about the
> third term of a triple using word embedding vectors with 6 components.
> Retrospectively, the components of these vectors could be interpreted as
> sensible features for capturing the structure of the domain (which was very
> conventional family relationships). For example, there was a three-valued
> feature for a person's generation and the interactions between features
> ensured that  the triple Victoria has-father ?  took the generation of
> Victoria and produced an answer that was of a higher generation because it
> understood that the relationship has-father requires this.  Of course, in
> complicated domains there will be huge numbers of regularities which will
> make conflicting predictions for the next word  but the consensus can still
> be fairly reliable. I believe that factoring the discrete symbolic
> information into a very large number of features and interactions IS
> intuitive understanding and that this is true for both brains and LLMs even
> though they may use different learning algorithms for arriving at these
> factorizations.   I am dismayed that so many people fall prey to the
> well-known human disposition to think that there is something special about
> people.
>
>
>
> Geoff
>
>
>
>
>
> On Mon, Mar 20, 2023 at 3:53 AM Paul Cisek <paul.cisek at umontreal.ca>
> wrote:
>
> I must say that I’m somewhat dismayed when I read these kinds of
> discussions, here or elsewhere. Sure, it’s understandable that many people
> are fooled into thinking that LLMs are intelligent, just like many people
> were fooled by Eliza and Eugene Goostman. Humans are predisposed into
> ascribing intention and purpose to events in the world, which helped them
> construct complex societies by (often correctly) interpreting the actions
> of other people around them. But this same predisposition also led them to
> believe that the volcano was angry when it erupted because they did
> something to offend the gods. Given how susceptible humans are to this
> false ascription of agency, it is not surprising that they get fooled when
> something acts in a complex way.
>
>
>
> But (most of) the people on this list know what’s under the hood! We know
> that LLMs are very good at pattern matching and completion, we know about
> the universal approximation theorem, we know that there is a lot of
> structure in the pattern of human-written text, and we know that humans are
> predisposed to ascribe meaning and intention even where there are none. We
> should therefore not be surprised that LLMs can produce text patterns that
> generalize well within-distribution but not so well out-of-distribution,
> and that when the former happens, people may be fooled into thinking they
> are speaking with a thinking being. Again, they were fooled by Eliza, and
> Eugene Goostman, and the Heider-Simmel illusion (ascribing emotion to
> animated triangles and circles)… and the rumblings of volcanos. But we know
> how LLMs and volcanos do what they do, and can explain their behavior
> without any additional assumptions (of thinking, or sentience, or
> whatever). So why add them?
>
>
>
> In a sense, we are like a bunch of professional magicians, who know where
> all of the little strings and hidden compartments are, and who know how we
> just redirected the audience’s attention to slip the card into our pocket…
> but then we are standing around backstage wondering: “Maybe there really is
> magic?”
>
>
>
> I think it’s not that machines have passed the Turing Test, but rather
> that we failed it.
>
>
>
> Paul Cisek
>
>
>
>
>
> *From:* Rothganger, Fredrick <frothga at sandia.gov>
> *Sent:* Thursday, March 16, 2023 11:39 AM
> *To:* connectionists at mailman.srv.cs.cmu.edu
> *Subject:* Connectionists: Can LLMs think?
>
>
>
> Noting the examples that have come up on this list over the last week,
> it's interesting that it takes some of the most brilliant AI researchers in
> the world to devise questions that break LLMs. Chatbots have always been
> able to fool some people some of the time, ever since ELIZA. But we now
> have systems that can fool a lot of people a lot of the time, and even the
> occasional expert who loses their perspective and comes to believe the
> system is sentient. LLMs have either already passed the classic Turning
> test, or are about to in the next generation.
>
>
>
> What does that mean exactly? Turing's expectation was that "the use of
> words and general educated opinion will have altered so much that one will
> be able to speak of machines thinking without expecting to be
> contradicted". The ongoing discussion here is an indication that we are
> approaching that threshold. For the average person, we've probably already
> passed it.
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20230324/4e85aff9/attachment-0001.html>