the borg

IBdaMann wrote:

Spongy Iris wrote:

IBdaMann wrote:Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

No. Into the Night already covered this. SIRI and ALEXA are regular internet searches, just using voice recognition as an interface instead of your keyboard. They combine internet searches with "settings" and the voice recognition interface can issue commands as well by recognizing verbs in your speech.

I want to give you a clear definition of AI as it is, de facto, in today's industry, but it defies any sort of definition that discerns it from "useful/helpful" software that is "really finely tuned." I am normally the first to complain when something is not clearly/unambiguously defined, but with AI, I can only describe it and use it as a buzzword. I know what it is and I can discuss it with others who know what it is, but I am at a loss to define it as it is. I can only describe it.

A good example is your ability to just know when a child/teenager is lying to you (which is actually a belief that he is lying). You don't think about it, you don't engage any philosophy and you don't follow any sort of algorithm. You just "know." If asked "how do you know", you might "chalk it up to experience" or just resign yourself to something like "it was obvious" but the fact remains that your brain simply gave you an answer, with an accompnaying degree of certainty/possibility of error.

When developing software to replicate "recognition" of this nature, e.g. hearing a voice and "recognizing" words of a language resulting in understanding a message, or seeing a video stream, "recognizing" a particular animal resulting in an understanding that it is hunting, etc.., one does not develop an algorithm in standard fashion. One creates a (now) standard "neural net" which is trained (refined) via machine learning and continuously improved (made more accurate).

AI was not always this way, and was not always viewed in this light. In fact, four/five decades ago, AI was "defined" in many very silly ways, with my favorite (funniest) being "If it fools a human into believing that it is intelligent, then it is therefore artificially intelligent."

AI is of value because it enables humans to extend human "recognition" into data that is normally unintelligible. If I were to show you a 3-D grid of 243,888,345 data points, it would not tell you anything whereas an AI system can discern thousands of major patterns, and patterns within patterns, and construct visualizations (pie charts, quad graphs, timelines, etc..) that bring all of those insights to you in a form you can understand.

The reason AI is really big today is that we have transitioned to a world of BIG DATA, and businesses find that they need to make decisions based on overwhelming data that is nothing but confusing ... to humans. So they turn the data over to mature AI systems that discern the patterns, provide the analyses, render projections and elucidate impacts to any anticipated or possible changes.

AI is used to help/aid/assist decision makers make decisions. AI is not an internet search or a voice command. AI is not a brilliant chess algorithm or an extremely helpful "assistant." AI brings human "recognition" and "understanding" to new data, often to BIG DATA, allowing decision makers to "see" and "understand" the new data for what it is, and not as unintelligible data.

IBdaMann wrote:

Into the Night wrote:Automated learning loops are as old as AI.

Perhaps, but AI itself has evolved tremendously and wasn't automated in its present form until relatively recently.

Into the Night wrote: One example is automated chess and checker games. Using a series of given goals for solid gameplay, and using simple results as a feedback loop, such chess and checker programs get better and better.

Games aren't AI. A program that is "really good" isn't AI nor is the act of improving software. If you aren't pointing to improving the accuracy of an underlying pattern recognition, you aren't really talking about AI. Yes, you are improving an algorithm, but AI is a different paradigm and isn't built in the form of an algorithm (yes, technically there is an algorithm, but that algorithm cannot be described as it can with, say, a chess program). Any attempt to describe/explain the logic of an AI application will be incomprehensible.

keepit wrote:
itn,
It was called alexa. I'm not sure what it meant by the phrase, "I'm still learning."

Spongy Iris wrote:

IBdaMann wrote:
Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

keepit wrote:
I don't know what the actual definition of AI is.

keepit wrote:
Now, that's something that i haven't seen in ai, a sense of humor, but i bet some computer has already shown a sense of humor.

keepit wrote:
Swan,
I remember having this same argument 60 years ago with a good friend of mine. I was surprised that he got very mad at me and we were never as good friends after that. I've learned since then that one should never lose a friendship over politics.
Anyway, i've read too much robert heinlein scifi from the 60's and the preoccupation with robots and their programming. Because of what i've read, i'm not surprised about the various levels of intelligence and superintelligence that has been speculated about.
So, the step ahead from that to computers making independent decisions beyond their programming seems like a real threat to me. I don't see how we're going tp stop it though. Computers, even if they can make decisions beyond their programming, can still have tunnel vision just like people.

Also, some humans might build dangerous computers just because they can. Then, the scenario outlined at the beginning of this thread, where ai decides to cause civilization to grind to a halt by turning all computers off, and then reactivates itself after civilization is greatly fragmented.
Also, robots can build new versions of themselves.

Swan wrote:

keepit wrote:
I don't know what the actual definition of AI is.

Now that makes sense, because the definition of ai depends upon the needs of the speaker. I define ai as cognitive awareness of self, not computer programs running scripts to fool simpletons

Swan wrote:

keepit wrote:
itn,
It was called alexa. I'm not sure what it meant by the phrase, "I'm still learning."

Alexa has zero intelligence; it runs computer scripts designed to find info from spoken commands not typed text commands and to fool suckers

Into the Night wrote:

Swan wrote:

keepit wrote:
I don't know what the actual definition of AI is.

Now that makes sense, because the definition of ai depends upon the needs of the speaker. I define ai as cognitive awareness of self, not computer programs running scripts to fool simpletons

Most computer programs do not run a script.

IBdaMann wrote:

Spongy Iris wrote:

IBdaMann wrote:Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

No. Into the Night already covered this. SIRI and ALEXA are regular internet searches, just using voice recognition as an interface instead of your keyboard. They combine internet searches with "settings" and the voice recognition interface can issue commands as well by recognizing verbs in your speech.

IBdaMann wrote:
I want to give you a clear definition of AI as it is, de facto, in today's industry, but it defies any sort of definition that discerns it from "useful/helpful" software that is "really finely tuned." I am normally the first to complain when something is not clearly/unambiguously defined, but with AI, I can only describe it and use it as a buzzword. I know what it is and I can discuss it with others who know what it is, but I am at a loss to define it as it is. I can only describe it.

IBdaMann wrote:
A good example is your ability to just know when a child/teenager is lying to you (which is actually a belief that he is lying). You don't think about it, you don't engage any philosophy and you don't follow any sort of algorithm. You just "know." If asked "how do you know", you might "chalk it up to experience" or just resign yourself to something like "it was obvious" but the fact remains that your brain simply gave you an answer, with an accompnaying degree of certainty/possibility of error.

IBdaMann wrote:
When developing software to replicate "recognition" of this nature, e.g. hearing a voice and "recognizing" words of a language resulting in understanding a message, or seeing a video stream, "recognizing" a particular animal resulting in an understanding that it is hunting, etc.., one does not develop an algorithm in standard fashion. One creates a (now) standard "neural net" which is trained (refined) via machine learning and continuously improved (made more accurate).

IBdaMann wrote:
AI was not always this way, and was not always viewed in this light.

IBdaMann wrote:
In fact, four/five decades ago, AI was "defined" in many very silly ways, with my favorite (funniest) being "If it fools a human into believing that it is intelligent, then it is therefore artificially intelligent."

IBdaMann wrote:
AI is of value because it enables humans to extend human "recognition" into data that is normally unintelligible. If I were to show you a 3-D grid of 243,888,345 data points, it would not tell you anything whereas an AI system can discern thousands of major patterns, and patterns within patterns, and construct visualizations (pie charts, quad graphs, timelines, etc..) that bring all of those insights to you in a form you can understand.

IBdaMann wrote:
The reason AI is really big today is that we have transitioned to a world of BIG DATA, and businesses find that they need to make decisions based on overwhelming data that is nothing but confusing ... to humans. So they turn the data over to mature AI systems that discern the patterns, provide the analyses, render projections and elucidate impacts to any anticipated or possible changes.

IBdaMann wrote:
AI is used to help/aid/assist decision makers make decisions. AI is not an internet search or a voice command.

IBdaMann wrote:
AI is not a brilliant chess algorithm or an extremely helpful "assistant."

IBdaMann wrote:
AI brings human "recognition" and "understanding" to new data,

IBdaMann wrote:
often to BIG DATA,

IBdaMann wrote:
allowing decision makers to "see" and "understand" the new data for what it is, and not as unintelligible data.

Into the Night wrote:

IBdaMann wrote:

Spongy Iris wrote:

IBdaMann wrote:Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

No. Into the Night already covered this. SIRI and ALEXA are regular internet searches, just using voice recognition as an interface instead of your keyboard. They combine internet searches with "settings" and the voice recognition interface can issue commands as well by recognizing verbs in your speech.

Incorrect. Both SIRI and the Alexa service use AI. This is how they handle their dictionaries and is how they recognize a spoken word as a command. This part of either system uses a learning loop to refine the recognition process.

Beyond that, cognition is by individually written programs, giving the service a usefulness. Typically, these are lambda functions written in some language such as Java or Python. That part is NOT AI, just a pile of useful programs and linked to the recognized words coming from the Alexa (or SIRI) service.

IBdaMann wrote:
I want to give you a clear definition of AI as it is, de facto, in today's industry, but it defies any sort of definition that discerns it from "useful/helpful" software that is "really finely tuned." I am normally the first to complain when something is not clearly/unambiguously defined, but with AI, I can only describe it and use it as a buzzword. I know what it is and I can discuss it with others who know what it is, but I am at a loss to define it as it is. I can only describe it.

There is a clear definition. AI is a program the makes use of a learning loop. That loop can be automated or manually handled, but it is a required feature of the program.

That's it. That's AI.

IBdaMann wrote:
A good example is your ability to just know when a child/teenager is lying to you (which is actually a belief that he is lying). You don't think about it, you don't engage any philosophy and you don't follow any sort of algorithm. You just "know." If asked "how do you know", you might "chalk it up to experience" or just resign yourself to something like "it was obvious" but the fact remains that your brain simply gave you an answer, with an accompnaying degree of certainty/possibility of error.

This is actually a good example, with some exceptions. Just like in AI, when the child makes an error you try to impose a correction. An AI program will mindlessly adapt to the correction. A child does not. He may or may not conform by his own choice.

IBdaMann wrote:
When developing software to replicate "recognition" of this nature, e.g. hearing a voice and "recognizing" words of a language resulting in understanding a message, or seeing a video stream, "recognizing" a particular animal resulting in an understanding that it is hunting, etc.., one does not develop an algorithm in standard fashion. One creates a (now) standard "neural net" which is trained (refined) via machine learning and continuously improved (made more accurate).

This sums it up pretty well. The resulting learned matrix is sometimes called a 'neural net', or a 'cortex'. This matrix is designed for rapid random access and scanning, giving AI it's speed available today.

IBdaMann wrote:
AI was not always this way, and was not always viewed in this light.

AI was always defined like this. Remember, other definitions were made by people not familiar with computers (i.e. News reporters and movie writers).

IBdaMann wrote:
In fact, four/five decades ago, AI was "defined" in many very silly ways, with my favorite (funniest) being "If it fools a human into believing that it is intelligent, then it is therefore artificially intelligent."

Yeah. I've seen them too. These wacky definitions were made by idiots unfamiliar with computers and programming.

IBdaMann wrote:
AI is of value because it enables humans to extend human "recognition" into data that is normally unintelligible. If I were to show you a 3-D grid of 243,888,345 data points, it would not tell you anything whereas an AI system can discern thousands of major patterns, and patterns within patterns, and construct visualizations (pie charts, quad graphs, timelines, etc..) that bring all of those insights to you in a form you can understand.

Yes. This is another application of AI. Like most AI, the initial conditions of what a 'pattern' even IS is directly coded into the program. Plotting the selected or summary data is nothing more than a plotting program, and does not involve AI at all.

IBdaMann wrote:
The reason AI is really big today is that we have transitioned to a world of BIG DATA, and businesses find that they need to make decisions based on overwhelming data that is nothing but confusing ... to humans. So they turn the data over to mature AI systems that discern the patterns, provide the analyses, render projections and elucidate impacts to any anticipated or possible changes.

Business have found that analyzing data in this way is a lot faster than hiring armies of data analysts, which used to perform this function. Because of this, it is now possible to analyze MORE data than ever before in the same amount of time.

IBdaMann wrote:
AI is used to help/aid/assist decision makers make decisions. AI is not an internet search or a voice command.

AI is used in voice recognition. It is also used in building search engine results. Google 'spiders' crawl the web, looking for new sites and changes to old sites, then send this raw data to Google. AI is used to correlate this data into rapid search results on demand. As people change what they are searching for, the learning loop adjusts what data should be available for more rapid access.

IBdaMann wrote:
AI is not a brilliant chess algorithm or an extremely helpful "assistant."

AI is used in chess and checker programs. In any of these types of programs, the rules of the game are coded into the program, and various goals to produce 'winning' combinations are also coded (such as in chess, gaining control of the center of the board). As people play the game using this code, a learning algorithm is used to adjust play to counter the moves of the opponent better. Over time, the program gets better and better and defeating various opponents.

IBdaMann wrote:
AI brings human "recognition" and "understanding" to new data,

AI is not human. However, it DOES allow humans to understand massive amounts of new data more quickly.

IBdaMann wrote:
often to BIG DATA,

Just what is 'big data'?

IBdaMann wrote:
allowing decision makers to "see" and "understand" the new data for what it is, and not as unintelligible data.

Yes. This is the purpose of such AI programs.
It was also the purpose of human data analysts, but the program is much faster than they are, making that position obsolete.

keepit wrote:
So, is this analysis of data, even on a massive scale going to be in the best interests of humans, and how would we know?\

The analysis in this sentence is a noun.

keepit wrote:
I don't think your guy understands my sentence very well.

keepit wrote:
So, is this analysis of data, even on a massive scale going to be in the best interests of humans, and how would we know?\

The analysis in this sentence is a noun.

keepit wrote:So, is this analysis of data, even on a massive scale going to be in the best interests of humans, and how would we know?

IBdaMann wrote:

keepit wrote:So, is this analysis of data, even on a massive scale going to be in the best interests of humans, and how would we know?

Currently, AI analyses are providing very beneficial results for SOME humans, e.g. Jeff Bezos (strategic positioning), Military Intelligence (red force anticipation), Sony (program selection/scheduling), etc ..

keepit wrote:
itn,
i used the word analysis as a noun, as i joe handed the analysis to bill.

keepit wrote:
Thanks Xadoman. Sometimes it seems like some people live on disinformation.

Swan wrote:

IBdaMann wrote:

keepit wrote:So, is this analysis of data, even on a massive scale going to be in the best interests of humans, and how would we know?

Currently, AI analyses are providing very beneficial results for SOME humans, e.g. Jeff Bezos (strategic positioning), Military Intelligence (red force anticipation), Sony (program selection/scheduling), etc ..

Again why would AI need a company?

keepit wrote:
itn,
i used the word analysis as a noun, as in, joe handed the analysis to bill.

keepit wrote:
Thanks Xadoman. Sometimes it seems like some people live on disinformation.

All English verbs (except to be) have five forms: base, past tense, past participle, present participle, and third-person singular.

Base Form (V1) analyse
Past Form (V2) analysed
Past Participle Form (V3) analysed
s / es/ es (V4) analyses
'ing' form (V5) analysing

Into the Night wrote:

Swan wrote:

IBdaMann wrote:

keepit wrote:So, is this analysis of data, even on a massive scale going to be in the best interests of humans, and how would we know?

Currently, AI analyses are providing very beneficial results for SOME humans, e.g. Jeff Bezos (strategic positioning), Military Intelligence (red force anticipation), Sony (program selection/scheduling), etc ..

Again why would AI need a company?

It doesn't. It is, however, expensive. Larger companies and government benefit most from it.

Xadoman wrote:
Also from google:

All English verbs (except to be) have five forms: base, past tense, past participle, present participle, and third-person singular.

For example:

Base Form (V1) analyse
Past Form (V2) analysed
Past Participle Form (V3) analysed
s / es/ es (V4) analyses
'ing' form (V5) analysing

Now lets say the word "alalysis" is indeed a verb. In which form it is? A base, past tense, past participle, present participle or third-person singular?
Also, just for clarification, could somebody write down all these five forms for the word "analysis"?

Xadoman wrote:
https://www.britannica.com/dictionary/analysis

Xadoman wrote:
Some examples from there:

- a scientific analysis of the data

Xadoman wrote:
- make/do/perform a chemical analysis of the soil

Xadoman wrote:
- a detailed analysis of the bone structure of horses

Xadoman wrote:
Analysis is a noun. Analyse is a verb. Simple as that.

make/do/perform a chemical analysis of the soil

Xadoman wrote:

make/do/perform a chemical analysis of the soil

make, do, perfrom - verbs

a - article

chemical - adjective

analysis - noun

of - preposition

the - article

soil - noun


Also, a simple way to distinguish between verbs and nouns. An article( a, an , the) goes with the noun. For example - an analysis( noun). "To" goes with the verb. For example - to analyse( verb).

IBdaMann wrote:

Xadoman wrote:
Also from google:

All English verbs (except to be) have five forms: base, past tense, past participle, present participle, and third-person singular.

For example:

Base Form (V1) analyse
Past Form (V2) analysed
Past Participle Form (V3) analysed
s / es/ es (V4) analyses
'ing' form (V5) analysing

Now lets say the word "alalysis" is indeed a verb. In which form it is? A base, past tense, past participle, present participle or third-person singular?
Also, just for clarification, could somebody write down all these five forms for the word "analysis"?

Xadoman, stop ... start over.

First, Google is saying that Google Assistant recognizes five verb forms. This is information for developers.

1. "Base" is the infinitive.
2. "Past Tense" is the preterite.
3. Past Participle
4. Present Participle
5. Third Person Singular

For the verb "to analyze" ... Google Assistant will have available

1. "analyze," "analyzing," and "to analyze"
2. "analyzed"
3. "analyzed"
4. "analyzing"
5. "analyzes"

HarveyH55 wrote:
I don't recall majoring in English, at a prestigious university, was a prerequisite for signing up for this site... Whining about spell, punctuation, and grammar, are deflections from the discussion topic. Someone who has truly mastered the English language, should have no problem making allowances for those who only managed the minimal required education. I was an English major, nor had much interest beyond a passing grade, but I have no problem comprehending through context.

As far as the thread topic goes... Never really got into the new Star Trek series. I liked the original series, but the newer ones are just a commercial product, selling liberal indoctrination.

IBdaMann wrote:

Spongy Iris wrote:

IBdaMann wrote:Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

No. Into the Night already covered this. SIRI and ALEXA are regular internet searches, just using voice recognition as an interface instead of your keyboard. They combine internet searches with "settings" and the voice recognition interface can issue commands as well by recognizing verbs in your speech.

A good example is your ability to just know when a child/teenager is lying to you (which is actually a belief that he is lying). You don't think about it, you don't engage any philosophy and you don't follow any sort of algorithm. You just "know." If asked "how do you know", you might "chalk it up to experience" or just resign yourself to something like "it was obvious" but the fact remains that your brain simply gave you an answer, with an accompnaying degree of certainty/possibility of error.

When developing software to replicate "recognition" of this nature, e.g. hearing a voice and "recognizing" words of a language resulting in understanding a message, or seeing a video stream, "recognizing" a particular animal resulting in an understanding that it is hunting, etc.., one does not develop an algorithm in standard fashion. One creates a (now) standard "neural net" which is trained (refined) via machine learning and continuously improved (made more accurate).

AI is of value because it enables humans to extend human "recognition" into data that is normally unintelligible. If I were to show you a 3-D grid of 243,888,345 data points, it would not tell you anything whereas an AI system can discern thousands of major patterns, and patterns within patterns, and construct visualizations (pie charts, quad graphs, timelines, etc..) that bring all of those insights to you in a form you can understand.

The reason AI is really big today is that we have transitioned to a world of BIG DATA, and businesses find that they need to make decisions based on overwhelming data that is nothing but confusing ... to humans. So they turn the data over to mature AI systems that discern the patterns, provide the analyses, render projections and elucidate impacts to any anticipated or possible changes.
.

Into the Night wrote:

Spongy Iris wrote:

IBdaMann wrote:
Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

The app itself is not, but the service they are connected to is using AI.

Spongy Iris wrote:

Into the Night wrote:

Spongy Iris wrote:

IBdaMann wrote:
Any attempt to describe/explain the logic of an AI application will be incomprehensible.

Is the SIRI or ALEXA app on my phone an example of AI?

The app itself is not, but the service they are connected to is using AI.

Ok here is another example. If I log onto YouTube and start watching political videos about Trump, and then almost all the videos on my feed start showing up as videos about Trump, is that AI?