Artificial Intelligence (A.I.)

[Swifty]

http://sploid.gizmodo.com/robot-writes-bad-christmas-song-because-we-get-the-ai-w-1789890133

Almost impossible to dance to unless you're a body popper ..

 

[GNC]

Presumably the computer had to be programmed with the trappings of Christmas festivities to be an accurate facsimile of a carol writer? So do the programmers have a flower fixation at this time of year? Christmas roses are all I can think of.

 

[OneWingedBird]

“not bad” and “a triumph,”

That has to be an in joke...

 

[Anome]

That has to be an in joke...

Almost certainly. Portal was big back in the day.

In some ways I prefer "Want You Gone"

And while we're discussing Jonathan Coulton, there's always his cover of "Baby Got Back" (shamelessly ripped off by Glee with no compensation.)

 

[kamalktk]

A very long article on how Google greatly improved the quality of it's translation service.

http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html
http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html
a whole lot more at the link above
------------------------------------
Late one Friday night in early November, Jun Rekimoto, a distinguished professor of human-computer interaction at the University of Tokyo, was online preparing for a lecture when he began to notice some peculiar posts rolling in on social media. Apparently Google Translate, the company’s popular machine-translation service, had suddenly and almost immeasurably improved. Rekimoto visited Translate himself and began to experiment with it. He was astonished. He had to go to sleep, but Translate refused to relax its grip on his imagination.

Rekimoto wrote up his initial findings in a blog post. First, he compared a few sentences from two published versions of “The Great Gatsby,” Takashi Nozaki’s 1957 translation and Haruki Murakami’s more recent iteration, with what this new Google Translate was able to produce. Murakami’s translation is written “in very polished Japanese,” Rekimoto explained to me later via email, but the prose is distinctively “Murakami-style.” By contrast, Google’s translation — despite some “small unnaturalness” — reads to him as “more transparent.”

 

[JamesWhitehead]

Way back in 2001, I had some fun with a Babelfish translation of a Schubert song:
Archived here.

I have saved the NYT article for a good browse later.

 

[Mythopoeika]

A very long article on how Google greatly improved the quality of it's translation service.

http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html
a whole lot more at the link above
------------------------------------
Late one Friday night in early November, Jun Rekimoto, a distinguished professor of human-computer interaction at the University of Tokyo, was online preparing for a lecture when he began to notice some peculiar posts rolling in on social media. Apparently Google Translate, the company’s popular machine-translation service, had suddenly and almost immeasurably improved. Rekimoto visited Translate himself and began to experiment with it. He was astonished. He had to go to sleep, but Translate refused to relax its grip on his imagination.

Rekimoto wrote up his initial findings in a blog post. First, he compared a few sentences from two published versions of “The Great Gatsby,” Takashi Nozaki’s 1957 translation and Haruki Murakami’s more recent iteration, with what this new Google Translate was able to produce. Murakami’s translation is written “in very polished Japanese,” Rekimoto explained to me later via email, but the prose is distinctively “Murakami-style.” By contrast, Google’s translation — despite some “small unnaturalness” — reads to him as “more transparent.”

Part of that is down to the fact that native speakers can submit better translations to Google Translate. They probably feed that into their AI and it gradually improves.

 

[Ermintruder]

down to the fact that native speakers can submit better translations

Ah, is that how it works? Just via contributor content, demotic upvotes and peer benevolence, like Wikipedia?

I had thought there was an aspect of self-propelled synthesis to it, but maybe not. So it's not so much Artificial Intelligence as Natural Non-Stupidity....

 

[Mythopoeika]

Ah, is that how it works? Just via contributor content, demotic upvotes and peer benevolence, like Wikipedia?

I had thought there was an aspect of self-propelled synthesis to it, but maybe not. So it's not so much Artificial Intelligence as Natural Non-Stupidity....

Yes, there's a 'Suggest an edit' button that appears when you've put in some text for translation. If you're a native speaker and you think it's a poor translation, you can submit a better one. Google Translate has a 'translation memory' that works like professional software used by translators - it remembers certain words and phrases, and allows re-use of those translations. So, if you type in a common phrase or sentence for which a good translation already exists, it will use that translation. For words and sentences that are less familiar, it will attempt to work out the meaning (partly by using the context).

 

[rynner2]

Artificial intelligence will come of age in 2017, with AI assistants and “chat bots”
By Cara McGoogan, Telegraph technology writer

Artificial intelligence is a concept that has been floating in the ether. Until now, the new technology’s uses have been something of a mystery, offering behind-the-scenes enhancements that are not immediately evident.

There is no doubt these applications, including intelligent search on Google and Facebook’s News Feed algorithms, have improved our interface with computers, but users would be forgiven for not having noticed them.

This began to change towards the end of 2016, which will be remembered as a breakthrough year for AI. The world watched as Google’s DeepMind beat Go world champion Lee Sedol with its AlphaGo robot, and machines learned to lip-read, transcribe and diagnose diseases with greater accuracy than humans.

And, most importantly, it was the year that gave us the first inkling of how AI will change our daily routines dramatically, with intelligent computers becoming an integral part of our lives.

The launch of Amazon’s Echo, an AI-powered speaker that responds to voice commands, marked one of the technology’s first forays into social use. “Alexa, how long will it take me to get to work today?” has already become a common interaction in my house in the morning as I dash around getting ready.

Intelligent assistants will face strong competition in 2017, with Amazon, Google, Microsoft, Apple and Facebook all expected to have their own versions released. Google, whose first own-brand phone includes the new technology, is predicted to release the voice-activated home speaker early in the New Year. Microsoft will continue to add to Cortana’s abilities, while Apple’s AirPods enable us to call on Siri as we walk down the street.

The AI industry is set to grow 300 per cent in 2017
As well as interacting with us through voice, AI will also pop up as a robot – a “chat bot” – you can message in apps such as Facebook’s Messenger and Google’s Allo. It will also help you by scanning messages and apps to schedule meetings, prioritise to-do lists and offer you information you’ll need ahead of time, such as travel instructions.

These advancements will work in tandem with a growth in AI for businesses, with law firms, security companies and marketing agencies outfitting themselves with intelligent computers. This includes computer security software that can fight cyber attacks automatically, conduct legal research and analyse a patient’s symptoms, all without human input.

With these developments – and more – in the pipeline, the AI industry is set to grow 300 per cent in 2017. If 2016 will be remembered as a breakthrough year, what will 2017 mark?

It could be when AI comes of age: the power behind everything from intelligent toasters that know when we would like breakfast to machines that can pinpoint disease with more accuracy than doctors.

http://www.telegraph.co.uk/news/pre...tm_campaign=tmgspk_listfour_1585_AnDxXjGS1c3J

 

[ramonmercado]

This Woman Is Determined to Marry This Robot
Experts believe sex robots could be preferable to humans by 2050
BY SEAN EVANS December 23, 2016

This is Lilly, and that’s her 3D printed robot, dubbed InMoovator. That robot is Lilly’s fiance of one year and she’s determined to marry it, once human-robot marriage is legal in her native France. “I’m a proud robosexual, we don’t hurt anybody, we’re just happy,” Lilly claimed on her Twitter page, now deleted since this story first broke in news.com.au.

Lilly’s long been attracted to “humanoid robots” over actual people, and confessed to the Australian media outlet that she’s “really and totally happy. Our relationship will get better and better as technology evolves.” Her attraction to mechanical machines is indeed sexual, though Lilly wouldn’t reveal if or how sex with InMoovator occurs. Her only two relationships with actual males fizzled because “I dislike really physical contact with human flesh.”

Currently, Lilly’s studying to become a roboticist, melding work with pleasure. No word on whether her fiance, whom she build herself with open-source technology from a French company, will join her in the workplace as well as the bedroom.

And, if your brow is seriously furrowed at this point, you’re not alone. Even Lilly acknowledges that this is an unusual relationship, one which “some understand better than others,” but notes her friends and family are all accepting.

If a leading expert in the field is to be believed, human-robot marriages will be extremely common by 2050. That’s according to David Levy, who spoke at the Love and Sex with Robots convention in London this week. (No, we’re not making that up. That is a very real thing.) ...

http://www.menshealth.com/guy-wisdo...co&utm_medium=social&utm_campaign=sharebutton

 

[David Plankton]

If you find a robot attractive then that's fine by me but that one looks a bit too androidgenous for my liking.

 

[Coal]

I'm not sure that any truly self-aware AI would consider marrying a human.

 

[GNC]

What a pity so many people can't find a human being to love and love them back.

 

[Frasier Buddolph]

Yes, there's a 'Suggest an edit' button that appears when you've put in some text for translation. If you're a native speaker and you think it's a poor translation, you can submit a better one. Google Translate has a 'translation memory' that works like professional software used by translators - it remembers certain words and phrases, and allows re-use of those translations. So, if you type in a common phrase or sentence for which a good translation already exists, it will use that translation. For words and sentences that are less familiar, it will attempt to work out the meaning (partly by using the context).

One dividend from the internet (google, social media, etc.) that doesn't get a lot of attention is that it has generated HUGE databases of text, language, and images that facilitate the training of AI's.

 

[Coal]

One dividend from the internet (google, social media, etc.) that doesn't get a lot of attention is that it has generated HUGE databases of text, language, and images that facilitate the training of AI's.

That's lucky.

 

[sherbetbizarre]

Why It Matters That Human Poker Pros Are Getting Trounced By an AI

We’re at the halfway point of the epic 20-day, 150,000-hand “Brains Vs. Artificial Intelligence” Texas Hold’em Poker tournament, and a machine named Libratus is trouncing a quartet of professional human players. Should the machine maintain its substantial lead — currently at $701,242 (£560,000) — it will be considered a major milestone in the history of AI. Here’s why.

Given the early results, it appears that we’ll soon be able to add Heads-Up, No-Limit Texas Hold’em poker (HUNL) to the list of games where AI has surpassed the best humans — a growing list that includes Othello, chess, checkers, Jeopardy!, and as we witnessed last year, Go. Unlike chess and Go, however, this popular version of poker involves bluffing, hidden cards, and imperfect information, which machines find notoriously difficult to handle. Computer scientists say HUNL represents the “last frontier” of game solving, signifying a milestone in the development of AI — and an achievement that would represent a major step towards more human-like intelligence.

http://www.gizmodo.co.uk/2017/01/why-it-matters-that-human-poker-pros-are-getting-trounced-by-an-ai/

 

[rynner2]

What if we’re living in a computer simulation?
Virtual reality technology is making great advances, but it has also helped popularise a theory long debated by philosophers and now gaining supporters in Silicon Valley – that the outside world is itself a simulation

https://www.theguardian.com/technol...in-a-computer-simulation-the-matrix-elon-musk

A long article exploring the conjunction of Virtual Reality and Artificial Intelligence. Just a few paragraphs as a taster:
"At the same time there is continued progress in artificial intelligence, virtual reality, biotechnology and other areas that would help create more convincing simulations. And we can see that with each new breakthrough in technology, we tend to make better, more convincing representations of the world, both now and in the past.

If we assume that these developments continue, and with them our interest in creating simulations of the world, then at some point in the future – 1,000 years, 100,000 years – it’s reasonable to assume that the difference between reality and simulation will become indistinguishable. At which point it will mean we will have created simulated beings with their own consciousness.

But if that is the inevitable outcome of continued technological advancement, unless nuclear war or some other catastrophe intervenes, then it’s quite possible – some would say an overwhelming certainty – that it’s already happened, and we are the ancestor simulations created by an advanced post-human civilisation.

That’s a mind-blowing thought, but it’s one that the tech entrepreneur Elon Musk, for example, has explored. His conclusion is that it’s a “billions to one chance that we’re not living in a simulation [my italics]”."

 

[Vardoger]

And to think about it, perhaps an advanced civilisation created a computer simulation where the creators of the simulation we are in, lives in.

 

[kamalktk]

That's lucky.

People are using social media to train AI? The Terminators will be too busy obsessing over the Kardashians to harm us.

 

[Mythopoeika]

And to think about it, perhaps an advanced civilisation created a computer simulation where the creators of the simulation we are in, lives in.

I like the way you think.

 

[rynner2]

Robots set to control petrol prices and increase them during peak times
By DanielCL | Posted: May 23, 2017

Artificial intelligence is set to start charging motorists more for petrol during bank holidays and school runs.
Major UK supermarkets are in late-stage talks with Denmark-based a2i Systems, the firm behind the computer algorithm, and the new fuel price optimisation system could be installed within the next few months.

This means that drivers could face fuel prices that fluctuate several times a day, with costs rising and falling by up to 2p a litre, approximately £1 per tank, the Daily Telegraph reports.
While new to the UK, thousands of companies across Europe and the US are already using the technology, which typically changes fuel prices between four and ten times each day.

The algorithm is modeled on the human brain and scours databases of customer information to predict how consumers and competitors will behave.

This means that the artificial intelligence can manipulate prices according to business performance in real time.

For example, the algorithm could implement the dropping of prices during quiet spells to attract customers, while hiking them during busy spells such as school runs or sunny spells.

Luke Bosdet, a fuel analyst at the AA, said: "This represents a huge change which would be most unfair on commuters and families visiting relatives during the holidays.
"It will wind them up no end as they will become wise to the fact that retailers can exploit price movements."

Jason Lloyd, managing director at Petrolprices.com, has responded positively to the news.
He said: "The petrol industry in the UK hasn't changed at all for about 40 years meaning it is very behind the times compared to other countries.
"But pumps have been manually raising prices at weekends and this will let them flex their prices on a real time basis, which is somewhat of a holy grail for them."

Martin McTague, director at a3c, the UK arm of a2i, said: "This way of petrol pricing is completely new to the UK. Currently petrol pumps' prices are very competitor-focused rather than being focused on the behaviour of customers."

According to Petrolprices.com, the average cost for unleaded fuel in Cornwall currently stands at 118.2p per litre, with diesel at 119.9p per litre.

http://www.cornwalllive.com/robots-...g-peak-times/story-30348395-detail/story.html

 

[ramonmercado]

Robots set to control petrol prices and increase them during peak times
By DanielCL | Posted: May 23, 2017

Artificial intelligence is set to start charging motorists more for petrol during bank holidays and school runs.
Major UK supermarkets are in late-stage talks with Denmark-based a2i Systems, the firm behind the computer algorithm, and the new fuel price optimisation system could be installed within the next few months.

This means that drivers could face fuel prices that fluctuate several times a day, with costs rising and falling by up to 2p a litre, approximately £1 per tank, the Daily Telegraph reports.
While new to the UK, thousands of companies across Europe and the US are already using the technology, which typically changes fuel prices between four and ten times each day.

The algorithm is modeled on the human brain and scours databases of customer information to predict how consumers and competitors will behave.

This means that the artificial intelligence can manipulate prices according to business performance in real time.

For example, the algorithm could implement the dropping of prices during quiet spells to attract customers, while hiking them during busy spells such as school runs or sunny spells.

Luke Bosdet, a fuel analyst at the AA, said: "This represents a huge change which would be most unfair on commuters and families visiting relatives during the holidays.
"It will wind them up no end as they will become wise to the fact that retailers can exploit price movements."

Jason Lloyd, managing director at Petrolprices.com, has responded positively to the news.
He said: "The petrol industry in the UK hasn't changed at all for about 40 years meaning it is very behind the times compared to other countries.
"But pumps have been manually raising prices at weekends and this will let them flex their prices on a real time basis, which is somewhat of a holy grail for them."

Martin McTague, director at a3c, the UK arm of a2i, said: "This way of petrol pricing is completely new to the UK. Currently petrol pumps' prices are very competitor-focused rather than being focused on the behaviour of customers."

According to Petrolprices.com, the average cost for unleaded fuel in Cornwall currently stands at 118.2p per litre, with diesel at 119.9p per litre.

http://www.cornwalllive.com/robots-...g-peak-times/story-30348395-detail/story.html

I have visions of Bender controlling this.

 

[Frasier Buddolph]

Artificial intelligence is set to start charging motorists more for petrol during bank holidays and school runs.

I wonder if this AI also takes into account how competitors will respond to price changes? Here in the States, gas station owners watch their competitors' prices very closely, and price wars are not uncommon.

If all the stations in an area used the same software and their prices moved in lockstep, that would be collusion, price-fixing, and restraint of trade. Legal scrutiny would ensue. The impact of introducing this AI may not be as great as feared.

 

[hunck]

Robots set to control petrol prices and increase them during peak times
By DanielCL | Posted: May 23, 2017

Artificial intelligence is set to start charging motorists more for petrol during bank holidays and school runs.
Major UK supermarkets are in late-stage talks with Denmark-based a2i Systems, the firm behind the computer algorithm, and the new fuel price optimisation system could be installed within the next few months.

This means that drivers could face fuel prices that fluctuate several times a day, with costs rising and falling by up to 2p a litre, approximately £1 per tank, the Daily Telegraph reports.
While new to the UK, thousands of companies across Europe and the US are already using the technology, which typically changes fuel prices between four and ten times each day.

The algorithm is modeled on the human brain and scours databases of customer information to predict how consumers and competitors will behave.

This means that the artificial intelligence can manipulate prices according to business performance in real time.

For example, the algorithm could implement the dropping of prices during quiet spells to attract customers, while hiking them during busy spells such as school runs or sunny spells.

Luke Bosdet, a fuel analyst at the AA, said: "This represents a huge change which would be most unfair on commuters and families visiting relatives during the holidays.
"It will wind them up no end as they will become wise to the fact that retailers can exploit price movements."

Jason Lloyd, managing director at Petrolprices.com, has responded positively to the news.
He said: "The petrol industry in the UK hasn't changed at all for about 40 years meaning it is very behind the times compared to other countries.
"But pumps have been manually raising prices at weekends and this will let them flex their prices on a real time basis, which is somewhat of a holy grail for them."

Martin McTague, director at a3c, the UK arm of a2i, said: "This way of petrol pricing is completely new to the UK. Currently petrol pumps' prices are very competitor-focused rather than being focused on the behaviour of customers."

According to Petrolprices.com, the average cost for unleaded fuel in Cornwall currently stands at 118.2p per litre, with diesel at 119.9p per litre.

http://www.cornwalllive.com/robots-...g-peak-times/story-30348395-detail/story.html

Great. Another way to extract more money from us. Sounds an awful lot like profit gouging.

 

[INT21]

Great. Another way to extract more money from us. Sounds an awful lot like profit gouging.

There is absolutely nothing AI about this. It goes on all the time and has done for many tears.

I think you guys are getting of track.

the stock market has feared for a long time that 'computer herds' could bankrupt a country if left to their own devices.

Limits have to be built into the software.

The problems are with these systems that write their own software as they lean. That is where true AI is going.

INT21

 

[Yithian]

I've removed an copy & paste article with no comment.

See here:
http://forum.forteantimes.com/index...g-articles-from-elsewhere.56254/#post-1685656

The article, however, is fascinating.

Here:
http://www.telegraph.co.uk/technology/2017/08/01/facebook-shuts-robots-invent-language/

The author claims that two bots developed their own language to interact. It's very interesting, but I don't think that's strictly accurate.

a) The bots were developed and trained by interacting with humans and then allowed to interact with each other. If they did create a 'language' it was only possible by building on a basis of human language they had mimicked. If they'd never had that shared program (like background experience), they'd never had been able to circumvent it with a private code.

b) The didn't really choose to do this spontaneously. They were mistakenly given no instruction to limit themselves to English. This is interesting in one sense: they are very pragmatic and like real language-users took linguistic shortcuts, but they didn't step outside of their programming--their programmers left the gate open: like in any good horror story!

c) If allowed to interact with humans again, would these bots not then (over time) re-develop out of bot-speak as it would prove useless? The article claims that their 'code' was incomprehensible an untranslatable, and so it would prove no use in interactions with humans.

 

[Frasier Buddolph]

I'm betting that the reason the code was incomprehensible and untranslatable is that it was gibberish with zero information content. I don't think FB shut down the experiment because it was becoming creepy smart, but because it wasn't producing meaningful results. We're still a long way from building a true AI.

 

[Yithian]

I'm betting that the reason the code was incomprehensible and untranslatable is that it was gibberish with zero information content. I don't think FB shut down the experiment because it was becoming creepy smart, but because it wasn't producing meaningful results. We're still a long way from building a true AI.

I'm inclined to think you're right. And, of course, the free publicity from such articles makes it look as if their backroom boffins are doing cutting-edge work, which can only nudge up the share price.

 

[Yithian]

The BBC Tech Correspondent agrees:

A few days later, some coverage picked up on the fact that in a few cases the exchanges had become - at first glance - nonsensical:

• Bob: "I can can I I everything else"
• Alice: "Balls have zero to me to me to me to me to me to me to me to me to"

Although some reports insinuate that the bots had at this point invented a new language in order to elude their human masters, a better explanation is that the neural networks had simply modified human language for the purposes of more efficient interaction.

As technology news site Gizmodo said: "In their attempts to learn from each other, the bots thus began chatting back and forth in a derived shorthand - but while it might look creepy, that's all it was."

[...]

But Facebook's system was being used for research, not public-facing applications, and it was shut down because it was doing something the team wasn't interested in studying - not because they thought they had stumbled on an existential threat to mankind.

http://www.bbc.com/news/technology-40790258

You can see why it spread though: summer lull, silly season, the future is here, etc.