Artificial intelligence

Se habla deScience!

Únase a LibraryThing para publicar.

Artificial intelligence

1scarper
Nov 19, 2011, 5:53am

Hi folks
Has anyone got a recommendation for a popular (ie not a text book) and up to date introduction to artificial intelligence?

2guido47
Nov 19, 2011, 6:20am

May I second that request,

I have read the emperors new mind and the sequel.
And was wondering if something of that same quality, with updated thinking, was around.

3DugsBooks
Nov 19, 2011, 11:12pm

I can post the obvious - I saw IBM's Watson on Jeopardy the the trivia game show. Here is a link to its official site:

http://www-03.ibm.com/innovation/us/watson/index.html

Some one is sure to write about that if it has not already been done. Apple's new IPhone4S with voice recognition and response it getting to the area I like. Imagine teaming the IPhone with a Watson, like having a conversation with a "Google search brain" no doubt. A more productive society?

http://www.apple.com/iphone/features/siri.html

From my casual reading a lot of people think autonomous satellites are the way of the future {see link}, it makes sense to me. Maybe software like that would have helped the Russian Mars moon project. Rats, that would have been neat to visit a moon of Mars.

http://www.physorg.com/news/2011-01-spacecraft.html

4timspalding
Editado: Nov 20, 2011, 1:12am

I'd be very interested to hear too. I read a number of books in the 80s and 90s, but nothing since.

5vy0123
Nov 20, 2011, 3:54am

There are talks on brains, minds, machines here. One of the talks explains why AI hasn't delivered. Marvin Minsky has an iTunes U course that refers to his Society of Mind first published in 1988. The speakers may have accessible webpages linking to books of interest published more recently.

6_Zoe_
Nov 20, 2011, 11:44am

Final Jeopardy is about Watson.

7andyl
Nov 20, 2011, 12:47pm

#3

I would just note that Siri isn't actually an iphone thing. It actually works on a big server. The iphone sends the recorded voice to the server and receives data back in return. Apple have chosen to link it to just the iphone 4S by requiring a unique ID (associated with the phone). Siri existed before the iphone and was bought up by Apple.

Of course to my mind Siri is a long way from real AI.

8DugsBooks
Nov 20, 2011, 6:34pm

Great links #5, I hope to view a few of MIT links soon.

#7 All the easier to hook up to a Watson perhaps? I guess A.I. will arrive in increments, unless one springs forth fully formed from the forehead of a "big blue" or other supercomputer. ;-)

9andyl
Nov 21, 2011, 4:10am

#8

Well if we are talking about strong AI, then I have my doubts that it will ever exist. Weak AI that will fool most of the people most of the time should be achievable and could be attacked in a piecemeal fashion.

10timspalding
Nov 21, 2011, 12:08pm

I'm still not seeing the basic book here—an entertaining survey along the lines of Machinery of the Mind: Inside the New Science of Artificial Intelligence, but written recently. Does one not exist?

11scarper
Nov 24, 2011, 4:35pm

I would have thought so, but maybe not? There's a new project for you Tim.

12timspalding
Nov 24, 2011, 9:28pm

The book that needs to be written would be sort of the "second coming of AI." After Machinery of the Mind artificial intelligence took a very rapid tumble in attention and prestige. It's come back in the last decade, rescued more by the availability of enormous new quantities of data on the web than anything.

13DugsBooks
Dic 13, 2011, 9:39pm

Just a note here, I saw an article on quantum computing at Extreme Tech - a must have for A.I.s in a lot of SF.

http://www.extremetech.com/extreme/108573-worlds-first-programmable-quantum-phot...

14guido47
Dic 13, 2011, 10:50pm

I wonder if it will follow "Moores law"?

15vy0123
Editado: Dic 26, 2011, 6:03am

By the time the ideas are collected in a book, they could be stale by about 3-years. Another place to look/join/ask is here.

The Emotion Machine is a more recent work by Minsky* who I understand was involved in imagining HAL for the film of 2001.

-- append

* author discusses consciousness and other topics

16vy0123
Dic 16, 2011, 8:21am

Perhaps a gifted writer can use the literature surveys of recent years and produce an entertaining basic read with funding if needed from kickstarter.

Another place to have a look is iRobot and another name is Rodney Brooks. Maybe you could get an interview and pointers.

17vy0123
Dic 16, 2011, 8:46am

#12. artificial intelligence took a very rapid tumble in attention and prestige

Maybe the practitioners became quants and hedge fund managers and the students abandoned their degrees to startup Google and the like.

18timspalding
Dic 16, 2011, 10:08pm

>17 vy0123:

Snort. Maybe so.

19vy0123
Dic 16, 2011, 11:44pm

20DugsBooks
Editado: Ene 28, 2012, 3:55pm

#14  "I wonder if it will follow 'Moores law'?"

Being highly unqualified to answer that question I will put my nonexistent reputation at risk and say , after scanning several articles on what a quantum computer is, that it sounds like they are using a lot of stuff to "contain the quantums"   ;-)   in that the article makes the "chips" sound rather large at the moment.

One advantage being that in a regular chip the state can be on or off, 0 or 1 value and in the quantum computer there are 4 spin types   edit: {make that 3, I think}  which a value can exhibit.  {My interpretation, someone with actual knowledge please feel free to correct}  I am guessing the structures containing the particles with quantum states and their logic gates can be further refined and reduced in
size.  One linked article in a post above states that the quantum computer is just a continuation of Moore's Law.

21vy0123
Dic 19, 2011, 11:23pm

There's Wired magazine issue 19-01.

22DugsBooks
Ene 26, 2012, 10:28pm

I just finished a chapter in Kaku's Physics of the Future ,published in 2011, titled "Future of AI" which had the following recomendations the Age of Spiritual Machines and The Singularity Is Near by Kurzweil . Kaku calls Kurzweil "the spokesperson for the singularity".

Kaku also quotes Mitch Kapor, founder of Lotus Development Corp as saying that the concept of singularity is "intelligent design for the IQ 140 people..." and is "driven by a religious impulse". I guess that leaves me saying those new computers sure look shiny ;-)

23DugsBooks
Editado: Feb 2, 2012, 10:47am

From Dilbert comic strip.


24vy0123
Nov 28, 2012, 5:07am

http://cser.org

An academic was reported to say to Elon Musk there's a 12% chance of life ending this century, one of the threats is a general A.I. It is interesting the software entrepreneur among philosopher and scientist driving the above initiative is Skype's co-founder.

25BogAl
Nov 28, 2012, 8:05am

>24 vy0123:

Coincidentally, there's a 12% chance that sometime this century the bank's automated phone service will actually recognize what my Midwestern girlfriend is saying to it.

26stellarexplorer
Nov 28, 2012, 1:17pm

>24 vy0123:. No wonder he's spearheading an effort to get people off Earth and onto Mars, where it may be safe.

http://www.wired.com/wiredscience/2012/11/elon-musk-mars-colony/

27guido47
Nov 28, 2012, 2:39pm

Having just 'acquired' almost all the 'Culture' books by Iain M Banks and no, I haven't yet read them, "...they just looks real pretty...", I do like the ideas of "strong AI." in SF. I recon. this is a direction, in SF, which most reminds me of my initial interest in the 50's/60's :-

28DugsBooks
Editado: Nov 30, 2012, 12:04am

#26 Great link Stellar, the most interesting article I have read on Musk. Being unburdened by not having very much engineering knowledge I still think my idea for a combination magnetically levitated and rail gun launch system located high on the S. American west coast could be a cheaper way to supply a "reusable" interplanetary "mother" ship that would shuttle between Earth & Mars etc. without needing to land anywhere. Kind of along the 2001 Space Odyssey lines.

Unless proven impossible on the drawing board it would seem a team like those that put together one of the large atom smashers could design the rail gun launcher.

::edit:: Maybe Musk's fact that fuel is only 2% of the cost of an escape velocity rocket is a factor in the equation for a different launch system.... come to think of it. But still I would like to see the idea mused over by some crack engineers.

::edit 2:: aha, the actual quote of Musk is "I tend to approach things from a physics framework. And physics teaches you to reason from first principles rather than by analogy. So I said, OK, let’s look at the first principles. What is a rocket made of? Aerospace-grade aluminum alloys, plus some titanium, copper, and carbon fiber. And then I asked, what is the value of those materials on the commodity market? It turned out that the materials cost of a rocket was around 2 percent of the typical price—which is a crazy ratio for a large mechanical product"

#25 The A.I. is just frustrating your girlfriend to make her more compliant - until she starts feeding in your credit card numbers ! ;-)

29vy0123
Nov 29, 2012, 1:06am

#28 EM has the hyperloop idea he plans to run numbers on. He believes the nano tubular elevator up is unlikely for various reasons, for one, now, you can't make a three legged stool out of the stuff. He's said he'd go to Mars once he's certain his companies will go on doing what he has in mind without him. His idea for re-orienting a column of water to shield occupants in spacecraft in the event of big solar flare makes sense.

On the six month trip to Mars, if there is an A.I. on board, I guess you want two of them, like the two sides of the brain.

Anyone have favorite books with spaceship A.I. they recommend ?

There's 2001

30stellarexplorer
Nov 29, 2012, 10:37am

I like the ships in Peter F Hamilton's The Reality Dysfunction, the Night's Dawn Trilogy.

31vy0123
Nov 29, 2012, 6:49pm

#30 Just imagine, sentient starships ( SS ) filled with artificial wombs for unwanted unborns shot in every direction. The SS search to locate a livable planet would be like a stripling sperm ( SS ) on a grand scale, traveling in the dark.

32DugsBooks
Editado: Jun 12, 2013, 1:50pm

I found this article interesting, apparently on topic and very recent; IBM packs 128TB of flash into brain-simulating supercomputer

Quote from the article:
"Computer scientists have long looked at brains for inspiration, and in recent decades, one avenue has been in self-learning neural networks that behave in some ways like a brain's teeming interconnected neurons. The Blue Brain project, though, isn't a neural network. Instead, it seeks to simulate the physical reality of real nerve cells, including their shape, size, and electrical behavior, IBM said."

Following one of the links you come across this 15 minute TED lecture by Henry Markram titled ; A Brain In A Supercomputer

http://www.ted.com/talks/henry_markram_supercomputing_the_brain_s_secrets.html

Quote from above lecture link:
"Henry Markram is director of Blue Brain, a supercomputing project that can model components of the mammalian brain to precise cellular detail -- and simulate their activity in 3D."

33DugsBooks
Editado: Abr 14, 2020, 6:31pm

"Artificial intelligence is evolving all by itself", new article in Science magazine. Can't be that technical, I can understand most of it.

First paragraph of article:

Artificial intelligence (AI) is evolving—literally. Researchers have created software that borrows concepts from Darwinian evolution, including “survival of the fittest,” to build AI programs that improve generation after generation without human input. The program replicated decades of AI research in a matter of days, and its designers think that one day, it could discover new approaches to AI.

https://www.sciencemag.org/news/2020/04/artificial-intelligence-evolving-all-its...

34stellarexplorer
Abr 14, 2020, 8:28pm

>33 DugsBooks: So is this good news, I wonder?

35pgmcc
Abr 15, 2020, 2:58am

>33 DugsBooks: Very interesting.

>34 stellarexplorer: We all remember Yul Brynner in Westworld.

36stellarexplorer
Abr 15, 2020, 4:11am

>35 pgmcc: I remember the tube behind his mouth that allowed him to consume liquids and maintain the image of hard-drinking SOB

37DugsBooks
Editado: Abr 15, 2020, 5:00pm

>34 stellarexplorer: I hope so. I have some medical issues I am being treated for and while Chemo, CarT and other techniques are improving it seems that something could wade through the masses of genetic data and find some correlation to predict the effectiveness of a treatment for a particular individual with a particular genome. That as opposed to hard won statistical data from clinical use with different patients - which works but with varying degrees of success as I understand it.

AI tweaking out molecules, biological pathways/mechanisms that might help thwart cancer or the current corono virus would be nice also. I read where some vaccine candidates were proposed using an A.I.

>35 pgmcc: Battlestar Galatica is being run in the USA on some of the "over the air" side band channels , which should remind us of safe guards that might be needed. ;-)

38pgmcc
Abr 15, 2020, 6:00pm

>37 DugsBooks: The real Battlestar Galatica or the remake?

39DugsBooks
Abr 15, 2020, 7:12pm

>38 pgmcc: The good one - the remake with Katee Sackhoff! IMOHO 🙂

40DugsBooks
Editado: Abr 29, 2020, 11:11am

A.I. can’t solve this: The coronavirus could be highlighting just how overhyped the industry is
An article at CNBC describing a sober evaluation of the current status of A.I. in addressing tough problems.

Quotes:
>Big AI firms like DeepMind, OpenAI and Facebook AI have remained relatively quiet throughout the crisis.

>AI has huge potential, but right now scientists are turning to tried and tested technologies instead.

>Of course, there are a few useful AI projects happening here and there.


41pgmcc
Abr 29, 2020, 12:31pm

My understanding is that AIs are programmed to "learn" (I use that word in a limited context) based on experience. Their decision making algorithms will be tuned by experience. That is a bit like forecasting based on only historical data. When there is a problem that comes along that turns all the traditional processes, practices and norms on their head the AI, which has been formatted on old data, will not be able to produce any appropriate responses without major external, i.e. human (for the time being), input.

42DugsBooks
Mayo 21, 2020, 11:34am

Microsoft Built One of the Most Powerful Supercomputers in the World to Develop Human-Like AI
https://www.extremetech.com/extreme/310806-microsoft-powerful-supercomputer-huma...

Interesting short article about Microsoft building some hardware specifically adapted for AI development. Quote from article:

"Microsoft and OpenAI announced a partnership last year to develop new artificial intelligence technologies, and Microsoft just revealed the first product of this deal: a massively powerful supercomputer."

43DugsBooks
Editado: Mayo 28, 2020, 10:55am

Eye-catching advances in some AI fields are not real
https://www.sciencemag.org/news/2020/05/eye-catching-advances-some-ai-fields-are...

A short article in Science magazine that provides some detail {a lot of which is over my head,terms can be looked up if not too lazy at the time, but very understandable to anyone with a software background I imagine} when comparing claims made by AI. Describing a published paper by some folks at MIT the article examines claims made by AI producers. The sifting process lauds some claims and lets a little air out of others.

44DugsBooks
Editado: Jul 24, 2020, 3:44pm

Tired of networking your Raspberry Pi's together as a dark web server and ready to move up in the tech world? Nvidia And University Of Florida Supercharge Education With AI Supercomputer, a quote from the article:

Nvidia and The University of Florida (UF) are in engaging in a unique public-private partnership that could result in the development of the largest university AI supercomputer in the United States, and perhaps the world. The project is anchored by a $25 million gift from UF alumnus and Nvidia co-founder, Chris Malachowsky along with a matching $25 million grant from Nvidia in the form of hardware, software, training, and services, as well as an additional $20 million from UF for data center upgrades.

Looks like they are promoting the use of AI across all disciplines.

https://www.forbes.com/sites/tiriasresearch/2020/07/21/nvidia-and-university-of-...

Another article on same subject:
https://www.hpcwire.com/2020/07/21/university-of-florida-to-deploy-world-class-a...

45proximity1
Editado: Ago 2, 2020, 9:11am

Artificial Intelligence: What Everyone Needs to Know

See also: The Neuroscience of Intelligence

Why this latter? Because, unless one understands more than just a seat-of-the-pants idea of the neurological basis of intelligence, one isn't prepared to grasp the enormous gulf which separates what's called machine-"intelligence" and what we mean by human intelligence.

46DugsBooks
Ago 1, 2020, 11:51pm

>45 proximity1: Looks interesting. I guess it is recent enough, published in 2016, to still be relevant :-) I will check some libraries when time allows.

47proximity1
Editado: Ago 2, 2020, 6:11am

of course it's recent enough. Four years old--things haven't and don't change that much on this topic and, really, they won't.

The technology can and of course shall get better and better, more realistic-looking. But no matter its progress, it isn't and it's never going to qualify as the crucial thing: the experience of a self-aware entity which not only reacts to stimulus but does this without and independent of any prior human intervention. That means that the intelligence arises spontaneously from a set of conditions which are only liable to but not certain to produce something recognizable as intelligent response to stimulus.

Contrived machines cannot, as a matter of principle, do that. They have no cognizance of an identity which is theirs as assumed rather than imposed.

48proximity1
Editado: Sep 10, 2020, 11:06am

The Guardian (London)

(Opinion)
| Artificial intelligence (AI)
"A robot wrote this entire article. Are you scared yet, human?"
by (Guardian editors misstating the by-line as belonging to A.I. Unit ) "GPT-3"

I.O.W.... bullshit and "click-bait".

Upshot, conclusion:

RE: "A robot wrote this entire article" (says The Guardian), or,

"Humans lie." And humans at The Guardian are "pants-on-fire" liars.

49DugsBooks
Editado: Ene 23, 11:31am

Deepmind solves 50-year-old ‘grand challenge’ with protein folding A.I.

I find this article interesting.

I remember people {chemists} talking about the difficulties in predicting protein/molecule folding over 30 years ago and attempts to use computers to aid with this. I hope this new knowledge helps to advance the techniques used to treat diseases - we sure need some better options with cancer.

::several tortuous rewordings::

50stellarexplorer
Ene 24, 10:52pm

>49 DugsBooks: I’d like to understand the strategy better, but this is very exciting!

51proximity1
Editado: Ene 25, 8:55am

>49 DugsBooks:


It would be much more interesting and helpful if, instead of posting a link to MS-NBC's report of the report, you posted a link to the relevant study (usually found in the news article). There's a standard form unique identifier --a "D.O.I." number--which is a reference used to identify academic studies. With that, interested readers could at least see if there's an
open-access version of the paper concerned at Cornell University's https://arxiv.org/

there is so much published work in this area that finding the paper referenced is difficult without the D.O.I number or the paper's author(s)' name(s) and release date.

E.g.

https://arxiv.org/search/?query=protein+folding+data+computer+analysis&searc...

52DugsBooks
Editado: Ene 27, 12:59pm

>51 proximity1: Duly noted and understood. I knew when posting that MS-NBC is not a great reference for research but that is where I became aware of the event. That, combined with being a little lazy left me to let anyone reading the post ferret out more detailed information.

Thanks for posting a procedure to do that!

53proximity1
Ene 27, 4:12pm


>53 proximity1:

SO, to read this, I have to go to MS-NBC, open a user's account, register my e-mail and other personal details--all to get an openly viewable link to this study which, when clicked on, displays the paper's URL?

No thanks.

Frankly, I don't want MS-NBC, or CNN or other aggregiously "Woke" organizations taking and storing my personal data just to access their press articles. If you'd like readers here to access and read about this study, you'd post a simple link to it.

54DugsBooks
Ene 27, 5:58pm

>53 proximity1: I don’t have to do any of that! I access the article and still have a “sign in “ button at the top right. I list some stocks I am interested in there so I may of have registered at one time or another but never receive emails from them. They have a paywall pro section for some articles but I haven’t joined. I will post several links to the same news as I have done in others & link to original published research if readily available in the future.

Thanks for pointing that out.

55proximity1
Ene 28, 5:01pm


>54 DugsBooks:



"This content is blocked because you are not allowing cookies.

"To view this content, click on Manage Cookie Preferences here or at the bottom of the page to allow all cookies."



I refused cookies. The article's version I was allowed to see included no link(s) to any study---though there may have been none, after all. This is simply an article pimping a claimed break-through by "Alphabet's" (Google-owned) Deep Mind.

So, apparently, not a lot to "see" there.

56DugsBooks
Ene 29, 12:17am

>55 proximity1: wandering off topic here but I respect your opinion on cookies etc. I always try to refuse cookies on sketchy sites ,afraid of malware, but it is obvious from margin ads which pop up that I am leaving a huge trail for spiders, bots etc.

57proximity1
Ene 29, 11:58am


It's alright. I think you got right to the point and saw and echo my own concerns. I don't (and can't) really know what's been excluded there from the view of those readers who refuse the cookies--that''s the angle, after all--but I do know that, interestingly, if there was one, there was no link to any study, independent or otherwise.

It's essentially a very promotionally-favorable article about the wonders of what they claim Deep Mind can accomplish.

Watson and Crick, building on and using essential work and contributions from Maurice Wilkins and Rosalind Franklin, were more modest in their claims of what they'd achieved.

58DugsBooks
Editado: Feb 15, 3:33pm

>57 proximity1: Ooops! CNBC just turned into a piece of crap today and exhibits all the lousy issues you describe. This had never happened to me in years of use and starting an "account" there did not improve matters. Using the app on an apple device still seems ok . I won't be linking to anything there again as you suggest. My apologies for being snarky & disregard all of the above. I sent them an email feedback. It is really too bad, they have some good articles on occasion.

edited in::

Farrah (CNBC)

Feb 15, 2021, 12:03 PM EST

Hi Doug,

Thank you for contacting CNBC Customer Care.

We appreciate you taking the time to report this issue to our team. Our team is aware of this issue and is actively working on resolving any problems as soon as possible. We will contact you when the issue is either corrected or a solution is found.

Thank you for your patience and please don't hesitate to reach out again with any additional questions or comments.

Sincerely,
Farrah
CNBC Customer Care
customercare@cnbc.com