Sergio Navega's
Selected Newsgroup Messages
In this page, I've compiled a list of selected messages that I posted in comp.ai, comp.ai.philosophy
and comp.ai.nat-lang in the last few years. If one can
keep up with the occasional nonsense uttered by kooks and crackpots (sometimes it is even
fun!), then you might find Usenet to be a fantastic place to exercise good reasoning and
argumentation. In reality, this is why I like Usenet: it is a place to test
yourself, to teach something to somebody, but very often to learn a lot. I
strongly recommend posting to Usenet, if you want to improve your argumentation abilities
and are not afraid of receiving strong criticism (or, eventually, plain insults).
The big question in Usenet is that it survives because of disagreement. If two people agree with each other, then their thread will extinguish very fast. But if there's a lot of strong opposing viewpoints being put forward, then you can expect a flood of messages. So when you post something, expect that somebody will reply to you with some kind of reproach. Don't take it too personally. However, you would be better off knowing what you're talking about.
What follows is a selection of some good contributions I wrote, in reverse chronological order. You may still query my viewpoints about any of these subjects: drop me a message at snavega@attglobal.net. (sizes shown only for messages greater than 10 k).
Re: Emotion in Robots (11 Jan 2000)
Intelligence is not created through logic, suitability of computers to mimick brain
processes.
Re: Case Based Reasoning - is this AI? (11 Jan 2000)
What is CBR (Case-Based Reasoning), why it is important, cognitive models of analogical
reasoning.
Re: arguments about humans and computers (10 Jan 2000) (51 k)
Rationalists versus Empiricists. Information processing strategy of the brain. Children's
first abstraction. College-level learning of symbols prior to sensory equivalents.
Analogical reasoning, binding, perception, genetics and intelligence, predispositions
versus innateness, domain specificity of language, genetic predispositions of emotions,
mirror neurons.
Re: Thinking without language? (20 Dec 1999)
Thoughts without linguistic counterparts, visual reasoning, Herbert Simon, Pylyshyn,
Kosslyn, innateness of language, 'semantics precedes syntax'.
Re: language (13 Dec 1999) (15 k)
What is grammar, developmental issues in children, universal grammars, what drives grammar
acquisition, DNA coding, language origin and evolution.
Re: Defect of Turing Test? (12 Nov 1999) (45 k)
Turing test, disembodied computers, subjective assessments of the TT, intelligence and
learning, CYC Project, predisposition to learn, creativity, blind humans talking about
color, concept development, creativity as necessary condition for intelligence, complexity
of learning, NFL, self-organization of visual cortex, random explorations, Copycat,
children learning, Intelligent Agents and their environments.
Re: A Major Stumbling Block (5 Nov 1999) (17 k)
Neural spikes, PET, fMRI, ERP (Event Related Potential), effect of grammatically and
semantically incoherent phrases on ERP, MMN (Mismatch Negativity).
Re: Cognitive concept shifts in human learning (5 Nov 1999)
Bibliography on concept formation and creativity.
Re: Rationality and Formalizability (4 Nov 1999) (103 k)
Formalization of the mind, theoretical vs empirical research, statistically derived
patterns, populations of neurons, stories and sequences of events, a two-level cognition,
creative ways to organize data, unsupervised learning, interaction with world, principle
of cognitive economy, complexity of learning, self-organization, patterns from meaningless
symbols, binding of sensory perceptions, the search for invariances, the origins of
curiosity, non-inductive origins of knowledge.
Re: More on neuroscience... (29 Oct 1999) (29 k)
Universal grammar, Chomsky, neuroscience methods, biological plausibility, references to
connectionist visions of language.
Re: Neurons and consciousness (30 Jul 1999) (38 k)
Perceptual systems, support for cognition, innate sensory layer, evolutionary adaptation
of senses.
Re: Induction and Evolution (27 Jul 1999) (32 k)
Induction, pattern recognition, random mutation, brains compared to databases, analogical
reasoning, novices and experts, problems of NNs, categorization, innate comparison
mechanisms.
Re: Neurons and consciousness (27 Jul 1999) (14 k)
What it takes to understand dictionary entries, recognition and generation.
Re: Neurons and consciousness (15 Jul 1999) (108 k)
Inductive origin of propositions, symbolic reasoning, innate vs learned things, feature
detectors, neural similarity detection, what is pattern, temporal sequences of spikes,
hierarchies of invariant aspects in patterns, equal vs similar spike trains, reinforcement
of similarity, one similarity judgement for all levels of cognition, elemental patterns in
the unconscious, brains good at "more data, faster processing", minimum
description length, recognition preceding generation, some cognitive experiments with
babies, Marcus/Pinker vs Elman/McClelland, surface aspects of language vs deep aspects,
grammars based on prime numbers, language as a dynamic system, brain plasticity,
competence without formal knowledge, innateness of language, Kanzi, symbolic patterns,
self-organization of brain, modularity doesn't prove innateness.
Re: Neurons and consciousness (14 Jul 1999) (79 k)
Origins of hypotheses, concept formation, judgements of similarity, perception, innate
mechanisms, behaviorism, effect of writing on civilization, symbolic abilities of humans,
scientific method, evolution of society, pseudo-science, prediction and explanation.
Re: Neurons and consciousness (13 Jul 1999) (23 k)
Induction, Popper, where do hypotheses come from, falsification, abduction, explanatory
power, creativity, mysticism.
Re: Neural Plasticity (2 Jul 1999)
This is my comment to Steven Pinker and John Skoyle's discussion on Evolutionary
Psychology mailing list regarding neural plasticity and learning.
Re: Creativity in Old Age (25 Jun 1999) (50 k)
Efficacy of creativity, causal models, defying spirit, old guys, search for the new.
Re: "imaging the brain while it
perceives" - Brooks (16 Jun 1999) (56 k)
Discussing definitions of intelligence, also an interesting example of an intelligent
washing machine.
Re: question (2 Jun 1999) (24 k)
Language as evolutionary advantage, problems with natural language understanding,
disembodied computers, Helen Keller, perception, syntax, semantics, meaning, etc.
On McCarthy, Logic and Cognition (28 May 1999)
(11 k)
Logic in AI, the frame problem, the need of a body for human-like intelligence.
Patrick Winston's DNA (20 May 1999)
A comment on Patrick Henry Winston's Keynote Address, given during AAAI-99.
Re: NN formats (17 May 1999) (20 k)
Comparing Artificial Neural Networks with biological counterparts. Questioning innateness
of language.
Re: NN formats (14 May 1999) (13 k)
On a confusion between Neural Networks and Semantic Networks and some disadvantages of the
latter.
What it takes to understand chinese (13 May 1999) (16 k)
Some comments on the Chinese Room thought experiment and the problem of rule-based
solutions to AI.
Re: Dr. Lenat theory (5 May 1999) (59 k)
On intelligence, common sense, first-order logic, the CYC project, pre-loaded knowledge,
behaving intelligently, perception, learning, etc.
The Baby in a Vat (4 May 1999) (169 k)
Inspired by Daniel Dennett's 'Brain in a Vat' thought-experiment, I devise another piece
to present some criticisms to John Searle's positions in the Chinese Room. The thread,
then, diverted to discuss symbolicist methods of cognition, hierarchical models of mind,
statistical methods, chaos/order, perception of patterns and regularities,
self-organizable networks, a machine to discover arithmetic operations, effect of
interaction on disambiguation, effect of reinforcement, alien communication with purely
syntactical signals.
Re: Seth's conjecture (29 Apr 1999)
AI as a collective enterprise, inner mechanisms versus external (social) ones.
Re: Mentifex musings on the comp.ai vote (28 Apr 1999)
Some advice on how to present one's ideas clearly and improve acceptance by peers.
On Reasonable Innate Knowledge (16 Apr 1999) (29 k)
What's innate and what's not, how this can affect our concept of learning algorithms.
Re: purely reactive intelligences (14 Apr 1999)
Intelligence, behaviorism, mental models.
Re: Is knowledge knowable? (8 Apr 1999) (27 k)
Novices, experts, analogy, deep concepts, understanding, abstractions, sensorimotor
grounding, Piaget.
Re: Biologically plausible mechanism for learning NL (7 Apr 1999)
Connectionist researchers on language acquisition.
Re: Letter to Another Editor (1 Apr 1999)
On McCarthy's intelligent thermometer.
Re: CYC (27 Mar 1999)
CYC, ontologies, knowledge-based systems.
Re: Just ... (27 Mar 1999)
An introduction to some philosophical doubts about AI, complexity, predictability.
Re: What is the place for Induction? (23 Mar 1999) (80 k)
Inductive forms of reasoning, perception, concepts, object recognition, feature detectors,
cognition in children, progressive specialization, etc.
Evidences of Mirror Neurons (17 Mar 1999)
Where I discuss recent studies of mirror neurons, activated when the animal acts or when
it sees another doing the same action.
Intermission: The HLUT brouhaha
During some weeks in March, I had been involved in a lively and interesting discussion about HLUTs (Huge Look-Up Tables). Normally, HLUTs are things that are dismissed as boring talk after no more than 10 minutes: they are irreal, impossible constructs with too many irrational suppositions, typical entertainment of armchair philosophers (which abound in comp.ai.philosophy). But on our way to find a "rationale" to try to prove that HLUTs can't work (which is, by itself, futile, because HLUTs work by definition) we find ourselves delving into a lot of important ideas related to the concept of natural intelligence. If you're nuts enough to read all this, start by the last post (first in chronological order), 'Re: An HLUT challenge' and go reading upward. My opinion about HLUTs changed dramatically at least two times in the course of the debates. But my previous concept of what it is to be an intelligent entity came out of this debate highly reinforced.
VOTE: Your HLUT Opinion is Valuable Here (12 Mar 1999) (68 k)
In which I proposed a poll to help in showing the opinions of the several sects in relation to HLUTs. It is funny that, among the ones who had courage to answer to the poll, few matches were found. What a heck of a thought-experiment!What I Think HLUTs Mean (15 Mar 1999) (81 k)
The Audio-CD and other experiments, the pointless idea of the HLUTs, again the quantum nature of our world, the question that HLUTs represent a sequel to the old, classical way of seeing the world as deterministic and mathematically exact, the fact that HLUTs mix mathematics with real world, with invalid results in both camps, and other topics.Why HLUTs are not intelligent (13 Mar 1999) (179 k)
Showing a bit more comprehension of what HLUTs mean, I propose an interesting way to see the question. This is the way to rule out deterministic-HLUTs. My oponents again stick to the "HLUTs know it all" idea, which is, obviously, a tautology, revealing the circular argument that is behind the whole idea of HLUTs. Oh, well, at least, my way of seeing things is more fun. I couldn't resist doing a trick with words (search for 'the church of the HLUT').
The Great Problem With HLUTs (11 Mar 1999) (51 k)
My first attempt to discredit HLUTs, based on a still myopic vision of mine of the breadth of the definition of HLUTs.
Start Here--v then read upward ^ (if you're brave!)
Re: An HLUT challenge (4 Mar 1999) (344 k)
This is the start of the HLUT story, a very large thread with lots of interesting stuff. HLUTs are idiotic things when seen by the "mathematical" side, but are very fun when seen by the "real world" side. Needless to say, I chose to put my viewpoint in the fun side and the discussion was long mostly because of that.
AI and Skepticism (1 Mar 1999) (121 k)
I propose to construct AI systems from a "skeptical engine"; then some relations
between AI and the scientific method are sketched, along with the utility of such a
machine to a researcher.
Re: Analog vs Digital (reposted) (28 Jan 1999)
Neurons as digital or analog processors, synchrony of populations of neurons.
Re: Analog vs Digital (reposted) (28 Jan 1999) (14 k)
An important message where I hypothesize the basic principles of an intelligent
architecture, with an example of vision perception made with simple symbolic structures.
Re: Rickert, Balter, lookup tables and intelligence (17 Feb 1999)
Two additional ideas for the impossibility of physical implementation of HLUTs. The big
issues, however, only on the posts above.
Re: Hahahaha (16 Feb 1999) (21 k)
Learning is not enough to characterize intelligence, also ability to be efficient in
several domains simultaneously.
Re: Hahahaha (15 Feb 1999) (16 k)
Natural and Artificial intelligence, intelligence also as a social (environmental)
phenomenon.
Re: Hahahaha (12 Feb 1999) (122 k)
In this large thread we start with Neural Nets in finance, AI evolving slowly, critique of
expert systems, AI compared to children, connectionism vs symbolicism, effect of hardware
on intelligence, critique of traditional takes on AI, statistical nature of input signals
and symbolic nature of abstractions, brain modularity and development, a tentative recipe
for the fundamental principles of intelligence, animal intelligence, keyboard as a sensory
input, Helen Keller, my interesting idea of artificial symbol grounding,
creativity, cognition from sensorimotor patterns, dynamic symbols, again the principles of
intelligence, biologically inspired NN models, neurons and intelligence, intelligence of
dogs as a question of degree, effect of language on intelligence and more.
Re: Hahahaha (12 Feb 1999) (34 k)
AI being more difficult than brain surgery, the concept of "intelligence
freezers", learning and intelligence, current AI applications, dumb software of
today, babies and CYC.
Re: EXCLUSIVE YOU (10 Feb 1999)
A simple neuroscientific vision of identity, V. S. Ramachandran's phantom limb.
Limits of Statistic Learning (5 Feb 1999) (22 k)
Type-1 and type-2 statistic learning problems, tractability and complexity of learning in
NN, innate feature detectors, evolution and learning.
Re: Intelligent behavior (30 Jan 1999) (50 k)
Senses defined by evolutionary pressures, native abilities versus artificially improved
ones, amazing auditory (perceptual) abilities.
On Rickert/Daryl's debate (29 Jan 1999) (34 k)
Some arguments against HLUTs (pointless, because you can't fight an ill-formed
mathematical definition), perceptual flexibility versus sensory flexibility,
"cocktail party" effect, evolution refining our senses, digitization.
Re: How to modify Copycat's slipnet (23 Jan 1999) (15 k)
On Douglas Hofstadter's Copycat mechanism, how to improve it, new problems to solve with
it.
Re: Letter to the editor (18 Jan 1999) (21 k)
On Moravec's future robots inheriting Earth, hardware evolution versus intelligence
evolution, accidents in discovery of new things favoring only the prepared mind.
Re: Letter to the editor (18 Jan 1999) (24 k)
Piaget's ideas, Lamarckism, Baldwin effect, genetic distribution.
Re: Are Semantic Links Enough? (12 Jan 1999) (49 k)
On the inadequacy of semantic links to represent important forms of knowledge, context of
conversation, brain as reducers of complexity, analogical reasoning, an interesting
contrived dialog between a human and an intelligent machine.
Re: Just a thought.... (28 Dec 1998) (14 k)
Top-down vs bottom-up approaches, problems of knowledge representation based on logic,
CYC.
Re: Question (28 Dec 1998) (10 k)
A suggestion for an interesting experiment in artificial life, with intelligent agents
learning where to look for food.
Re: Just a thought.... (26 Dec 1998) (14 k)
Dehaene's book "The Number Sense", innate knowledge in humans, the importance of
the Baldwin effect in the evolution of our society.
Re: Just a thought.... (23 Dec 1998)
AI as a multidisciplinary enterprise, copying functional principles of the brain's
operation.
Re: AI & Creativity (10 Dec 1998) (10 k)
Creativity in children, criticism to conventional teaching, knowledge and creativity,
understanding.
Best Names in AI, CogSci (3 Dec 1998) (43 k)
A selection of names of researchers and thinkers of importance to the field of AI,
Cognitive Science and Philosophy.
Re: Brain rewires itself (1 Dec 1998)
Visual and auditory improvement through plasticity.
Re: Can artificial intelligence really exist? (30 Nov 1998)
(48 k)
On Searle's Chinese Room and a different interpretation, revising GOFAI, fundamental
principles of AI, Dreyfus, Kosslyn, Pylyshyn, Fodor.
Re: Methods of machine learning? (18 Nov 1998)
A short list of some methods applied in machine learning.
Intelligence: Interaction is Essential (11 Nov 1998) (44 k)
Interactivist vision of intelligence, causal models, categorization, disambiguation,
intractability of learning, experimentation, deaf-blind people.
Technical Reports and Publications
Intelliwise Seminars (in Portuguese)
Messages in the Cognitive Science Mailing List (in Portuguese!)
Last updated in 11-set-2003