News for hackers. Stuff that matters.
Re:Facebook’s AI accidentally created its own language by merricksb on Tuesday June 20, 2017 @08:56AM (#14594217)
Discussed 5 days ago: (107, 16 comments)

https://news.ycombinator.com/item?id=14554101

[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by dump121 on Tuesday June 20, 2017 @08:47AM (#14594161)
Seems the AIs have taken personality of the developer and also their liking to create new language and frameworks.
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by edejong on Tuesday June 20, 2017 @08:20AM (#14593971)
More background with a nice article: https://s3.amazonaws.com/end-to-end-negotiator/end-to-end-ne...
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by TXV on Tuesday June 20, 2017 @08:59AM (#14594230)
Unfortunately neither the OP's article nor this one explain anything about the AI's own language. For me this is enough to make the thread a clickbait. However the paper points out the more interesting fact that the model learnt how to deceive in order to achieve its goal. I think it would be hard to demonstrate that the behavior is really deceit and not a result of more "boring" features of the training examples, like let's say - dialogue length, limited use of negations (no, don't, can't, ...) - or a biased input.
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by crsv on Tuesday June 20, 2017 @08:42AM (#14594119)
Seems like a sizzle article with no steak. Some anonymous quotes, no real evidence presented.
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by nxsynonym on Tuesday June 20, 2017 @09:01AM (#14594248)
Without any concrete examples its hard to say... but I wouldn't call what they say the AI are doing creating a new language. At most they seem to be going off script, which is creating a new dialog - but a "language" is click-baity and sensationalized.

Call me when the AI bots are talking to each other in a way we can't understand.

[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by neuronDen on Tuesday June 20, 2017 @08:23AM (#14593987)
Wrote a blog post a couple of days back pointing what's wrong with this kind of articles: https://rajat503.github.io/blog/post/ai-in-perspective/
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by a13n on Tuesday June 20, 2017 @08:47AM (#14594156)
It's really hard to read your article when your text is justified rather than left aligned. I immediately gave up.
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by neuronDen on Tuesday June 20, 2017 @08:52AM (#14594186)
Fixed it, thanks!
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by scribu on Tuesday June 20, 2017 @08:59AM (#14594232)
The title is a gross exageration.

If you read the actual paper, you'l find that it's not a new language, but merely "meaningful novel sentences" in English.

That is, instead of merely reusing sentences found in the training data, the model generates new ones, some of the time.

[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by thedangler on Tuesday June 20, 2017 @08:56AM (#14594219)
When this is indeed true, the creators of the AI should make it so that they can't write software so that they don't put themselves out of a job :)
[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by mywittyname on Tuesday June 20, 2017 @08:59AM (#14594235)
They'll still be need for human AI mechanics when things break down.

/s

[ Parent | on HN ]
Re:Facebook’s AI accidentally created its own language by digi_owl on Tuesday June 20, 2017 @08:31AM (#14594056)
Paging cstross...
[ Parent | on HN ]