N.Wells
Posts: 1836
Joined: Oct. 2005
|
| Quote (GaryGaulin @ June 22 2014,13:21) | | Quote (N.Wells @ June 22 2014,12:58) | No, that's worse.
Scientific precision cannot happen without linguistic precision.
I had noticed that you added text output, presumably in case some future-generation Watson communicates via text on a screen. That is stretching beyond recognition your point about motor control, but it doesn't matter much because the much larger issue is that your "motor" and "control" aspects still do not cover such things as thinking out how to explain a difficult concept, deciding what is wrong with a sentence, choosing an appropriate word, figuring out a chain of logic, naming a concept, devising an explanation, identifying a theory, or re-assessing your life. While excluding these things from the purview of "intelligence" may well explain your behavior, it won't impress anyone who holds to a more standard definition of intelligence. |
The definition and rest of theory includes all you mentioned. You only have a strawman argument. |
Strawman? Incorporates all my complaints? No and no.
| Quote | | Intelligence is defined as the ability to learn by trial and error (or to self-program), and is qualified as being intelligent by its four requirement systematics containing: 1; body (or modeling platform) with muscles (or motors, linear motor speaker, text output) to control, 2; sensory addressed memory to store motor actions, 3; one or more confidence (hedonic system) levels to gauge motor action failure or success, 4; guesses motor actions whenever something new is first encountered or confidence in a memory action sufficiently decreases. |
1) "Intelligence is defined here as the ability to learn by trial and error (or to self-program) and is qualified as being ......." That's three different things.
2) In addition to learning by trial and error, learning can also happen by analysis of the situation, by being taught, and by learning from the experience of others, among other ways. Are these to be excluded from intelligence?
3) "qualified": Poor word choice, applied contrary to its meaning: a behavior qualifies as intelligent (for example if it involved analysis and decision-making), or "his exam results qualified him for the job", not "intelligence is qualified by its attributes" or "he was qualified by his exam results".
4) "intelligence... is qualified as being intelligent": that's a tautology.
5) "requirement systematics": that is gobbledygook.
6) "requirements": Those four things are not defining characteristics or qualifications needed for something to meet the definition of "intelligence": making plans for your life does not require any of them.
7 & 8) ": 1;" and ", 2;"
The colon is unnecessary, but you can have it if you insist, but the semi-colon is an abomination in a list with commas as major separators. Use "; 2) " or ", 2)" or ", (2)" or some format like that.
9) “body (or modeling platform) with muscles (or motors, linear motor speaker, text output) to control”
Well, that’s better, but it’s still awkward. Even expressed as “a body with muscles to control” would still be awkward. In any case, stop interleaving the sets of nouns: e.g. say, “1) a body with muscles or a modelling platform with motors". If you wanted to say, “Intelligence probably took a giant leap forward when bodies developed muscles that would be much more useful under active control than when operated by simple reflexes”, no one would disagree.
.
10) “Platform”: You really don’t mean modelling platform. Visual Studio or Visual Basic is your modelling platform, or see http://vimeo.com/formmod....atform. See also https://salilab.org/imp............imp You just mean a computer model.
11) “to control”: intelligence does not have to have something to control to be intelligence - again, planning your life, or merely fondly remembering breakfast.
12) “text output”: Text output is not in the same categories as muscles and motors. You are making nonsense of your own arguments here.
13) “sensory addressed memory”. That would be “sensory-addressed memory”, except that
14) “sensory addressed memory” doesn’t actually mean anything. You probably mean the ability to address or recall sensory memories.
15) “store motor actions”: Poorly written: Things that store motor actions are capacitors, springs, flywheels, and the like. You probably mean "capability of remembering prior responses". Note that your statements excludes plants, etc., let alone molecules from having intelligence.
16 to 20) “3; one or more confidence (hedonic system) levels to gauge motor action failure or success”
16) confidence is not an hedonic system, so your in-apposition phrase is inapposite.
17) Intelligence requires a confidence level? No way. ("Confidence" and "confidence levels" is not the same as "a way to evaluate confidence".)
18) Intelligence doesn’t even require confidence. Can you remember your breakfast? Where does confidence enter into that? You are trying to say that one of the defining characteristics of intelligence is being able to assess levels of confidence in possible actions based on evaluation of outcomes in prior experience.
19) “Gauge”: What NoName said.
20) “motor action failure or success” Too many nouns in a row.
Confidence is a misnomer, and motor actions are a red herring: you perhaps want to argue that intelligence requires the ability to remember prior experiences, to evaluate the success or failure of those outcomes, and to learn from those lessons and apply them to current problems.
21-25) “guesses motor actions whenever something new is first encountered or confidence in a memory action sufficiently decreases”
21) Your four points are 1) [a] body or modelling platform, 2) a memory, 3) a way of evaluating prior successes and failures, and 4) “guesses” [as a verb]. Your construction is non-parallel, and you want a capability for guessing, not just “guesses”
22) “new” and “first” are redundant: just pick one.
23) "Guess": what NoName said
24) It’s not “confidence in a memory action”. What the hell is a memory action? You are trying to talk about confidence in the accuracy of a memory, or confidence that the remembered action will be helpful in the present situation, and your are conflating the two into a meaningless phrase.
25) Intelligence is not involved in making a guess after you’ve decided that you are out of intelligent options. It may well be a better strategy than doing nothing and it will be programmed into AI systems, but “making a guess” is not a required property for something to be intelligent.
26 through infinity) Again, your four criteria exclude instinctive behaviors and reflexes, at least at the level of the individual, and they exclude anything involving plants and non-motile organisms, and they rule out “molecular intelligence” and “cellular intelligence”. Evolution by natural selection lets a population "learn" in a sense, but no confidence or gauging or intelligence is involved: the population becomes adapted because genomes experience differential reproductive success. Natural selection and evolution by natural selection are well documented; “molecular and cellular intelligence” are not.
|
