Re: [Vo]:Article about Artificial Intelligence in NYT
I wrote: Unfortunately for the crickets, they cannot tell one another apart, and > they cannot tell the difference between a cricket and a plastic model of > one. So, naturalists who wanted to give a male cricket an inferiority > complex engaged in ritual combat with him using a plastic model of a > cricket. They did this over and over again with the same plastic model. The > poor guy-cricket did not realize he was fighting the same dummy cricket > every time. > My point being, that barely qualifies as conscious. Not as I defined it: "Awareness of surroundings. Some ability to make choices . . ." It is more like a set of complicated hard-wired reactions. The cricket mistakes a plastic object for another cricket. Its perceptions and awareness of the surroundings are very crude, compared to a bird or mouse. A self-driving car probably has a better "mental model" of the surrounding environment than the cricket does. Maybe not . . . you can easily fool a self-driving car with a two-dimensional cutout model of a pedestrian. I saw a video of that the other day. The car may not be A.I. based. I wouldn't know. - Jed
Re: [Vo]:Article about Artificial Intelligence in NYT
> There are an infinite number of steps between each level. > In the Universe of all material phenomena in general (i.e., so-called 'meta-physics' being pure Idealist wankerism), its development must necessarily be open-ended and emergent. How could it be any other way. So there will always be _further_ possible levels beyond any already achieved. By any entity or collection of entities.
Re: [Vo]:Article about Artificial Intelligence in NYT
a.ashfield wrote: > I think it follows we will have a "conscious" computer even earlier than > Kurzweil forecasts. I think there are various levels of intelligence. Roughly speaking here are three points on the spectrum: 1. Thinking. You can make a case that even guppies and earthworms do this. 2. Conscious. Awareness of surroundings. Some ability to make choices, rather than purely instinct driving, hard-wired brain functions. I expect that mice are conscious. In this article it was estimated that present-day artificial intelligence computers have roughly as many virtual synapses as a mouse brain has. 3. Sentient, or self-aware. At the lowest level, this means knowing that you are an animal and an object in the real world. There is no doubt that apes and other intelligent creatures have this. At the zoo in Boston, when you take a picture of a chimpanzee with a digital camera, it will pose and then demand to see the back of your camera. Especially males do this, according to my daughter, who is studying biology. My guess is that computers are somewhere between 1 and 2. They have probably not achieved 2 because people who design computers are not trying to achieve this at present. Perhaps consciousness will emerge on its own as a meta-phenomenon. There are an infinite number of steps between each level. There are various mental achievements. For example, male crickets are capable of fighting for domination, which is sophisticated behavior. It is impressive for such a small brain. Surely, this is a form of thinking, even if it is mainly instinct driven. Unfortunately for the crickets, they cannot tell one another apart, and they cannot tell the difference between a cricket and a plastic model of one. So, naturalists who wanted to give a male cricket an inferiority complex engaged in ritual combat with him using a plastic model of a cricket. They did this over and over again with the same plastic model. The poor guy-cricket did not realize he was fighting the same dummy cricket every time. Apparently this sapped his male hormone supply, a.k.a. precious bodily fluids. - Jed
[Vo]:starting list of LENR mysteries
http://egooutpeters.blogspot.ro/2016/12/dec-16-starting-list-of-lenr-mysteries.html peter -- Dr. Peter Gluck Cluj, Romania http://egooutpeters.blogspot.com
[Vo]:A road map to Interstellar Space Travel
https://arxiv.org/abs/1604.01356
Re: [Vo]:Article about Artificial Intelligence in NYT
Jed, Thank you for posting the link to that most interesting article. For some years now I had thought that consciousness was just the brain making an image of one's surroundings and examining that image. I see now that is unnecessary, that the brain examines the actual image in real time and doesn't have to make a new 3D image fifteen times a second and look at that. The basic concept remains though, that consciousness is simply the act of examining the image together with inputs from the other sensors. I think it follows we will have a "conscious" computer even earlier than Kurzweil forecasts. AA On 12/15/2016 5:47 PM, Jed Rothwell wrote: There was a long article about artificial intelligence (AI) in the New York Times: http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html It was pretty good, with some technical detail. You will find more detail in recent Sci. Am. article by two of the leading people in the field. I can't find it on line . . . There has been an important breakthrough with neural networks. They have been around for decades, going back to the 1950s I think. The difference is they are now many orders of magnitude larger, and they are multi-level, with the output of one network connected to the input of another. This was the technique that led to a computer beating the world champion in go. The NYT reports that Google has applied this to their translation software, resulting in dramatic improvements. In a few months, the quality of the translations improved as much as it did in 10 years with the older technology. The article quotes an example. This sentence in Spanish by Borges: Uno no es lo que es por lo que escribe, sino por lo que ha leído. The old Google translate system rendered this: One is not what is for what he writes, but for what he has read. The new one: You are not what you write, but what you have read. I ran some Japanese and some English text through the latest Google translate. This is mainly text that I translated myself. The new Google translate is remarkable. A little unnerving. Because, you see, if you run some of my translated essays you will see that I took liberties, adding information I thought would help a native speaker understand. These are not literal or exact translations. Since I wrote the original text myself, I am allowed to to that. But, I also did it with Mike McKubre's paper. I am a little worried that someone may call me out on it! Sooner or later, Google's computers will be getting in touch with me, calling me out. Google sells a gadget that sits in the room listening to your conversations, awaiting your commands, the Google Home: https://madeby.google.com/home/ I can see the day coming when the Google Home speaker will blare out: "ROTHWELL! Get over here. What is the meaning of this?!? McKubre wrote 'I was tasked' and you have it: 'the conference organizers asked me to . . .' We are now in the process of reviewing all of your work going back to 1998, which will henceforth be called Calendar Year 1 of Our Lord Google." http://lenr-canr.org/acrobat/McKubreMCHcoldfusionb.pdf http://lenr-canr.org/acrobat/McKubreMCHcoldfusionc.pdf - Jed
[Vo]:Google
I am also amazed by Google. They provide, free of charge, their appcompiler Android Studio to anyone; even me. Iwas able as an older retired, skills out of date, engineer, to figure it out. This was done without human instruction. I thenpublished apps into the global market. Isold an app in Brazil and one in Turkey. I know not to whom. If I can do this, so can millions more. Google, with the effort, has harnessed theproductive and artistic effort of the entire human civilization. The barriers to entry or low. Zero cost of entry. You must, however, be able understand the Java language and know how to use an XML file. I am not a computer engineer but and electrical engineer and I was able to do this. On the cold fusion front someone has been going through all of my references in detail. Perhaps it is nothing but maybe it is something. I know this is pompous but my book "Energy Cold Fusion and Antigravity" is of an epic proportion. I will be around long after the shrunken atom, and heavy neutrons, are gone. Maybe that story is beginning to unfold now. Frank Znidarsic
Re: [Vo]:Article about Artificial Intelligence in NYT
-Original Message- From: Jack Cole To: vortex-l Sent: Fri, Dec 16, 2016 7:10 am Subject: Re: [Vo]:Article about Artificial Intelligence in NYT It is rather remarkable. A recent story along these lines is that it can even translate between two untrained languages. For example, if the system has been trained to translate between English and Spanish and English and Portuguese, then it can reasonably translate between Spanish and Portuguese even though it has not been trained. https://techcrunch.com/2016/11/22/googles-ai-translation-tool-seems-to-have-invented-its-own-secret-internal-language/ On Thu, Dec 15, 2016 at 4:48 PM Jed Rothwell wrote: There was a long article about artificial intelligence (AI) in the New York Times: http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html It was pretty good, with some technical detail. You will find more detail in recent Sci. Am. article by two of the leading people in the field. I can't find it on line . . . There has been an important breakthrough with neural networks. They have been around for decades, going back to the 1950s I think. The difference is they are now many orders of magnitude larger, and they are multi-level, with the output of one network connected to the input of another. This was the technique that led to a computer beating the world champion in go. The NYT reports that Google has applied this to their translation software, resulting in dramatic improvements. In a few months, the quality of the translations improved as much as it did in 10 years with the older technology. The article quotes an example. This sentence in Spanish by Borges: Uno no es lo que es por lo que escribe, sino por lo que ha leído. The old Google translate system rendered this: One is not what is for what he writes, but for what he has read. The new one: You are not what you write, but what you have read. I ran some Japanese and some English text through the latest Google translate. This is mainly text that I translated myself. The new Google translate is remarkable. A little unnerving. Because, you see, if you run some of my translated essays you will see that I took liberties, adding information I thought would help a native speaker understand. These are not literal or exact translations. Since I wrote the original text myself, I am allowed to to that. But, I also did it with Mike McKubre's paper. I am a little worried that someone may call me out on it! Sooner or later, Google's computers will be getting in touch with me, calling me out. Google sells a gadget that sits in the room listening to your conversations, awaiting your commands, the Google Home: https://madeby.google.com/home/ I can see the day coming when the Google Home speaker will blare out: "ROTHWELL! Get over here. What is the meaning of this?!? McKubre wrote 'I was tasked' and you have it: 'the conference organizers asked me to . . .' We are now in the process of reviewing all of your work going back to 1998, which will henceforth be called Calendar Year 1 of Our Lord Google." http://lenr-canr.org/acrobat/McKubreMCHcoldfusionb.pdf http://lenr-canr.org/acrobat/McKubreMCHcoldfusionc.pdf - Jed
Re: [Vo]:Article about Artificial Intelligence in NYT
Jack Cole wrote: > . . . A recent story along these lines is that it can even translate > between two untrained languages. > > For example, if the system has been trained to translate between English > and Spanish and English and Portuguese, then it can reasonably translate > between Spanish and Portuguese even though it has not been trained. > > https://techcrunch.com/2016/11/22/googles-ai-translation- > tool-seems-to-have-invented-its-own-secret-internal-language/ > That's amazing. But it is not the best test they could have run. They translated from Japanese into Korean. Japanese and Korean are similar in many ways. This is like going from Spanish into Italian. They should try it from Japanese to Russian, or Japanese to Navajo. - Jed
Re: [Vo]:Article about Artificial Intelligence in NYT
It is rather remarkable. A recent story along these lines is that it can even translate between two untrained languages. For example, if the system has been trained to translate between English and Spanish and English and Portuguese, then it can reasonably translate between Spanish and Portuguese even though it has not been trained. https://techcrunch.com/2016/11/22/googles-ai-translation-tool-seems-to-have-invented-its-own-secret-internal-language/ On Thu, Dec 15, 2016 at 4:48 PM Jed Rothwell wrote: > There was a long article about artificial intelligence (AI) in the New > York Times: > > http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html > > It was pretty good, with some technical detail. You will find more detail > in recent Sci. Am. article by two of the leading people in the field. I > can't find it on line . . . > > There has been an important breakthrough with neural networks. They have > been around for decades, going back to the 1950s I think. The difference is > they are now many orders of magnitude larger, and they are multi-level, > with the output of one network connected to the input of another. > > This was the technique that led to a computer beating the world champion > in go. The NYT reports that Google has applied this to their translation > software, resulting in dramatic improvements. In a few months, the quality > of the translations improved as much as it did in 10 years with the older > technology. The article quotes an example. This sentence in Spanish by > Borges: > > Uno no es lo que es por lo que escribe, sino por lo que ha leído. > > The old Google translate system rendered this: > > One is not what is for what he writes, but for what he has read. > > The new one: > > You are not what you write, but what you have read. > > > I ran some Japanese and some English text through the latest Google > translate. This is mainly text that I translated myself. The new Google > translate is remarkable. A little unnerving. Because, you see, if you run > some of my translated essays you will see that I took liberties, adding > information I thought would help a native speaker understand. These are not > literal or exact translations. Since I wrote the original text myself, I am > allowed to to that. But, I also did it with Mike McKubre's paper. I am a > little worried that someone may call me out on it! Sooner or later, > Google's computers will be getting in touch with me, calling me out. Google > sells a gadget that sits in the room listening to your conversations, > awaiting your commands, the Google Home: > > https://madeby.google.com/home/ > > I can see the day coming when the Google Home speaker will blare out: > "ROTHWELL! Get over here. What is the meaning of this?!? McKubre wrote 'I > was tasked' and you have it: 'the conference organizers asked me to . . .' > We are now in the process of reviewing all of your work going back to 1998, > which will henceforth be called Calendar Year 1 of Our Lord Google." > > http://lenr-canr.org/acrobat/McKubreMCHcoldfusionb.pdf > > http://lenr-canr.org/acrobat/McKubreMCHcoldfusionc.pdf > > - Jed > >