In Define data earlier than you speak about it, neurosurgeon Michael Egnor interviewed engineering prof Robert J. Marks on the way in which data, not matter, shapes our world (October 28, 2021). In the primary portion, Egnor and Marks mentioned questions like: Why do two similar snowflakes appear extra significant than one snowflake? Now they flip to the connection between data and creativity. Is creativity a perform of extra data? Or is there extra to it?
This portion begins at 10:46 min. A partial transcript and notes, Show Notes, and Additional Resources observe.
Michael Egnor
Michael Egnor: How does organic data differ from data in nonliving issues?
Robert J. Marks: I don’t know if it does… I do imagine after latest examine that the thoughts could be very completely different from the bodily a part of the mind. So there’s data that happens exterior to the mind. In phrases of simply the bodily, materialistic definition, most data can be utilized to measure what’s within the organic entity.
We can speak about, nevertheless, creativity and the place the concept of creativity comes from — the creation of data. And that’s outdoors of naturalistic or data processes.
Michael Egnor: Thomas Aquinas, following Aristotle, outlined residing issues as issues that try for their very own perfection. He felt that was what distinguished residing issues from nonliving issues. A rock doesn’t on a regular basis get up within the morning and attempt to be a greater rock. Whereas residing issues to a higher or lesser levels of success, attempt to make themselves higher at what they do. They eat, they relaxation, they work together with nature, they do issues to make themselves even higher examples of what they’re.
It would appear to me which may relate to the distinction between data in non-living and residing issues. Thhe data in residing issues is directed to ends. It’s directed to functions that you just don’t see in nonliving issues in the identical means.
Robert J. Marks: I might undoubtedly agree with that it. It does prove that, with the intention to do the development that you just’re speaking about, there must be a level of creativity.
Robert J. Marks
This is among the issues that we argue quite a bit about in synthetic intelligence. Will synthetic intelligence ever be inventive? And I preserve that synthetic intelligence won’t ever be inventive, it’ll by no means perceive. And at the moment it has no frequent sense…
Michael Egnor: Sure. I imply, I’ve at all times considered synthetic intelligence as only a illustration of human intelligence. And that, within the sense the time period synthetic intelligence is an oxymoron. If it’s synthetic, it’s not intelligence. And it’s so intelligence have to be human. And all of the intelligence that’s in computer systems and pc applications and machines, is all human intelligence that’s represented in these gadgets.
Robert J. Marks: It is. In truth, what you talked about is precisely the take a look at that Selmer Bringsjord, a professor at Rensselaer, utilized in his take a look at for whether or not or not synthetic intelligence can be inventive. His take a look at was that does the pc program do one thing which is outdoors the reason or the intent of the programmer? And to this point no synthetic intelligence has completed this.
Note: The first one who is thought to understand that synthetic intelligence is just not inventive was pc pioneer Ada Lovelace (1815–1852), an affiliate of Charles Babbage (1791–1871). Selmer Bringsjord’s Lovelace take a look at for AI creativity is called for her.
Ada Lovelace
Robert J. Marks: Now now we have stunning outcomes from synthetic intelligence. AlphaGo, that beat Lee Sedol within the Asian sport of Go — probably the most troublesome board sport of all — at one level it did the stunning transfer. And all people went, “Oh, that’s unimaginable. This was inventive.” No, that was not creativity. AlphaGo was skilled to play go. And that’s precisely what AlphaGo did. It was enjoying go, higher than human beings. But we see this on a regular basis…
Note: Tech thinker George Gilder notes in his latest guide, Gaming AI, that the explanation AI sweeps the board in video games like go and poker is that, in video games, the map is the territory: The guidelines at all times work. The motive AI tends to do a lot worse in, say, drugs is that in drugs, the map is just not the territory. When dealing within the virtually infinitely extra advanced actual world amongst human beings not constrained by such guidelines, qualities resembling creativity are important.
Michael Egnor: So what particularly is the relevance of data to biology? And why is data so attention-grabbing when it’s utilized to a examine of residing issues?
Robert J. Marks: Well… it’s past the reason of materialism. We hear quite a bit in the present day about Ray Kurzweil’s Singularity: AI writes higher AI that writes higher AI and fairly quickly, the AI — it’s claimed — goes to have intelligence equal to that of a human being.
Well, in fact, if now we have AI writing higher AI, that’s past the unique intent of the pc programmer, then now we have creativity don’t we? And creativity is past the flexibility of synthetic intelligence. So Ray Kurzweil’s concept isn’t going to work.
There’s additionally the thought that sometime we can have one thing which is named synthetic basic intelligence, which will likely be intelligence on the human degree, together with issues like creativity, understanding and sentience. And no this all requires creativity by way of the pc program, and that may by no means occur.
I’ve a colleague, Roman Yampolskiy on the University of Louisville. Amid all this speak about synthetic basic intelligence, he posted a put up on social media on April Fool’s Day in 2016. Let me learn it to you:
Google simply introduced main layoffs of programmers future software program growth and updates will likely be completed largely through recursive self enchancment by evolving deep neural networks.
In different phrases, he stated that this [artificial general intelligence] AGI had been achieved. And this allowed all people to be laid off at Google as a result of they weren’t wanted anymore.
Roman Yampolskiy
Robert J. Marks: And it was a terrific joke. But on his social media, he bought a bunch of likes. But additionally curiously, a number of requests from journalists got here in they usually stated they wish to speak to him about this excellent growing story. So to non-experts, the joke was not apparent. But I believe to anyone that does synthetic intelligence, it was clearly an unimaginable joke. And I believe it’s humor that illustrates this concept that no, we’re by no means going to have the type of pc producing software program that generates higher and higher AI. Isn’t going to occur.
Note: Quite a few paradoxes come up with respect to AI and creativity. Creativity doesn’t observe computational guidelines. Programmers can’t write applications which might be extra inventive than they themselves are. And, as mathematician Gregory Chaitin notes, as soon as one thing is lowered to a components a pc can use, it isn’t inventive any extra, by definition. And the idea of synthetic intelligence itself designing even higher synthetic intelligence is likewise problematic.
Michael Egnor: Right. There’s a pc scientist in California named Judea Pearl, who I discover to be a captivating man. Over the previous a number of many years, he’s actually pioneered within the area of causal evaluation — the flexibility to ascribe quantitatively, utilizing arithmetic, causal inferences in processes, which traditionally has been very exhausting to do. Statistical analyses have been capable of infer correlation, however they weren’t capable of infer causation.
He is making an attempt to determine methods to enable machines to make causal inference. Because to have machines which might be price utilizing, which have any type of what seems to be autonomy, the machine has to have the ability to infer trigger… versus easy correlations.
Gary Smith
Robert J. Marks: Of course, the those who do phrase mining and pure language processing are large into issues resembling correlation. And correlation is simple to do. My good friend Gary Smith, a professor of economics at Pomona College, stated that is going to be very harmful within the space of knowledge mining. He stated as a result of we’re going to provide you with inferences that don’t have anything to do with one another.
There’s a terrific web site known as Spurious Correlations. They present how, in large information, you will get correlations between completely unrelated causes:
Michael Egnor: Machines can crunch numbers, and do issues that don’t present perception. But to get perception into what causes what, I believe, requires a human being.A very good instance [of a spurious correlation] is that people who smoke often have yellowing of their fingertips as a result of they’re holding cigarettes. And they’re additionally predisposed to get most cancers. There’s no query that yellowing of your fingertips correlates with having lung most cancers but it surely doesn’t imply that yellow fingertips trigger lung most cancers. You must get the causal arrows proper. Yellow fingertips and lung most cancers are each attributable to a typical issue, which is smoking, however they don’t trigger one another. News, “Yellow Fingers Do Not Cause Lung Cancer” at Mind Matters News (December 10, 2020)
Robert J. Marks: Establishing causation to this point goes to require human intervention. Yes. So yeah, I agree with you.
Michael Egnor: It’s type of attention-grabbing as a result of the inference to causation appears to be leap past odd algorithmic statistical evaluation of one thing. And it’s a leap, I believe that solely human beings could make… And there’s every kind of errors made in medical analysis. Because it’s assumed that correlates suggest trigger which they don’t essentially.
Robert J. Marks: And that is attention-grabbing, as a result of a lot of the journals that settle for papers based mostly on statistics are solely into the correlations. They require an R worth of such and such, which means that the information corresponds to a excessive diploma of correlation. But the issue is, in fact, we do have these spurious correlations… And this has led to the conclusion of 1 researcher that as much as 90% of the papers revealed within the literature which might be based mostly on statistics are flawed.
Michael Egnor: I might are inclined to disagree. I believe it’s a lot increased than 90%. I believe that’s an actual underestimate the variety of papers which might be rubbish.
Robert J. Marks: Michael, I used to be able to argue with you. It seems, okay.
Michael Egnor: Yeah 90% could be very conservative. Yeah.
Robert J. Marks: Which is de facto superb. And that’s the explanation that in the present day… Coffee is nice for you, espresso is dangerous for you. You hear all of those fleeting research which might be reported within the information as gospel, and it may possibly have a horrible impact, it may possibly make you paranoid. And so I attempt to ignore all of those since Gary Smith pointed this out to me. It’s actually horrible…
I keep in mind the same story in regards to the ice cream consumption and homicide charge to New York City — that the ice cream charge would enhance after which the homicide charge would enhance. And after folks bought bored with consuming ice cream, they went out and killed one another? They have been each simply associated to the rise in temperature.
Michael Egnor: Well, I used to be a resident in neurosurgery in Miami through the drug wars. And so we’d get gunshot wounds to the pinnacle coming into the ER continually. Except they might at all times cease when it rained. And Miami has fairly a little bit of rain. So once we would have an hour or two of rain, the ER would simply go fully quiet. Nobody would are available. And then the solar would come out after which folks would shoot one another once more. And it was fascinating, however folks wouldn’t shoot one another throughout rain.
Robert J. Marks: So the causation can be that good climate causes folks to kill one another.
Michael Egnor: Yeah, or that there’s one thing about being moist that protects you from gunshot wounds…
Next: Does Mount Rushmore don’t have any extra data than Mount Fuji?
Here’s the earlier episode within the sequence:
How data turns into every thing, together with life. Without the data that holds us collectively, we’d simply be mud floating across the room. As pc engineer Robert J. Marks explains, our DNA is essentially digital, not analog, in the way it retains us being what we’re.
You may want to learn: How data realism subverts materialism Within informational realism, what defines issues is their capability for speaking or exchanging data with different issues.
Show Notes
00:00:09 | Introducing Dr. Robert J. Marks00:01:02 | What is data?00:06:42 | Exact representations of data00:08:22 | A system with minimal information00:09:31 | Information in nature00:10:46 | Comparing organic data and data in non-living things00:11:32 | Creation of information00:12:53 | Will synthetic intelligence ever be inventive?00:17:40 | Correlation vs. causation00:24:22 | Mount Rushmore vs. Mount Fuji00:26:32 | Specified complexity00:29:49 | How does a statue of Abraham Lincoln differ from Abraham Lincoln himself?00:37:21 | Achieving goals00:38:26 | Robots enhancing themselves00:43:13 | Bias and concealment in synthetic intelligence00:44:42 | Mimetic contagion00:50:14 | Dangers of synthetic intelligence00:54:01| The position of data in AI evolutionary computing01:00:15| The Dead Man Syndrome01:02:46 | Randomness requires data and intelligence01:08:58 | Scientific critics of Intelligent Design01:09:40 | The controversy between Darwinian concept and ID theory01:15:07 | The Anthropic Principle
Additional Resources
Robert J. Marks at Discovery.orgMichael Egnor at Discovery.orgClaude Shannon at Encyclopædia BritannicaAndrey Kolmogorov at WikipediaSpurious Correlations websiteChapter 7 of: R.J. Marks II, W.A. Dembski, W. Ewert, Introduction to Evolutionary Informatics, (World Scientific, Singapore, 2017).Winston Ewert, William A. Dembski and Robert J. Marks II “Algorithmic Specified Complexity within the Game of Life,” IEEE Transactions on Systems, Man and Cybernetics: Systems, Volume 45, Issue 4, April 2015, pp. 584-594.Winston Ewert, William A. Dembski and Robert J. Marks II “On the Improbability of Algorithmically Specified Complexity,” Proceedings of the 2013 IEEE forty fifth Southeastern Symposium on Systems Theory (SSST), March 11, 2013, pp. 68-70Winston Ewert, William A. Dembski, Robert J. Marks II “Measuring significant data in photographs: algorithmic specified complexity,” IET Computer Vision, 2015, Vol. 9, #6, pp. 884-894
Podcast Transcript Download
