I am old enough to remember three A.I. booms and two intervening "winters". The first boom was picking all the low hanging fruit like playing checkers, moving blocks, solving word problems. The power of computers in the 60s and 70s was pitiful. Then the easy stuff was done for while.
The second boom was expert systems and logic super computers. Those systems never worked that well and A.I. went into long sleep.
Now its supercomputer data mining and greatly improved neural networks.
Back in 2005 I suggested that it would be interesting to build a system that would play any game without being given directions (that blog post has since disappeared from the web). This, and not image recognition, would mean intelligence, I thought at the time: you can't truly tell a donkey from a dog without having some general knowledge about the world donkeys and dogs (and your AI) are living in.
The academic image recognition machine though seems unstoppable and yes it does seem to improve over time. I honestly don't know what the limits of "dumb" image recognition are in terms of quality, but calling it AI still doesn't make sense to me.
AGI is the idea of creating general intelligence. Which almost certainly requires some reinforcement learning (the ways dogs and humans learn, by having a goal and an internal reward system).
Deepmind's Atari player is based on deep reinforcement learning, where increasing the score represents a reward:
I used to believe the moving goalpost idea, that AI is anything that isn't yet possible. I now disagree.
I saved myself a lot of energy by avoiding entirely the question "what counts as AI?" by switching to the question, "what counts as AGI?" which is a term with a clearer threshold:
> but calling it AI still doesn't make sense to me.
That's the core problem of AI, no matter what progress is made, it's instantly called not AI anymore while the goalpost of what AI is is continually being pushed out to not this. The real issue of course is what don't how intelligence actually works so it's impossible to set a fixed goalpost of when AI is truly achieved.
There isn't a lot in the world of AI that seems worthy of the I. Image recognition though, while not a part of our conscious reasoning is a very strong part of our brains. It continues to advance because it is immediately profitable.
That's actually starting to happen. Deepmind built an AI that can learn to beat the best human players on many atari games. After just a few hours of playing the game and learning. And of course it uses all that advanced image recognition stuff. That has always been the hard part. The actual game playing part is just a simple reinforcement learning neural network that is put on top of it.
The reason it's AI is because it isn't specific to speech recognition. Deep learning is very general. The same algorithms work just as well at speech recognition, or translation, or controlling robots, etc. Image recognition is just the most popular application.
This is true. The reason there's a scene in the movie, 2001, in which HAL plays chess is that at the time it was thought that playing chess well required real human intelligence.
But as soon as chess programs got good, we all took them for granted.
Image recognition requires AI. It used to be believed that it was simple. A famous AI researcher in the 50's once sent a bunch of grad students to solve it over the summer. They then started to realize just how complex and impossible the task was.
60 years later, we have finally made general purpose learning algorithms, vaguely inspired by the brain, which are just powerful enough to do it. And because they are general purpose, they can also do many other things as well. Everything from speech recognition, to translating sentences, or even controlling robots. Image recognition is just one of many benchmarks that can be used to measure progress.
Image processing appears pretty central to human thought hence phrases like 'throw some light on the problem' and 'get the full picture', if you see what I mean.
2700 deep learning projects at Google... What are they doing, besides the obvious? What's a good ball-park estimate of the number of "projects at Google"?
2700 doesn't strike me as remotely crazy. I'm guessing that some of those projects all serve the same goal, e.g., for their speech system, they have: acoustic modeling; language modeling; named-entity recognition; intent classification; domain classification; grapheme-to-phoneme conversion; language detection; wake-word detection. This ignores other stuff that happens around speech (for example, I know they were using a CRF to label different types of numbers for training their spoken-to-written form converters, which AFAIK are still using WFSTs, although at this point I wouldn't be shocked if both of those systems were converted to DNNs). So let's take an estimate of 10 DNNs for their speech systems. Per language, so make that 200 DNNs to support 20 languages. This ignores that they have separate models for YouTube, voice search (one model for on-device and a cloud-side model), voicemail.
Their machine translation system probably has a similar # of DNNs, and there you have to deal with language pairs, rather than single languages. Let's call it another 400.
That's two side-projects. Then you pull in query prediction, driverless cars, all kinds of infrastructure modeling, spam detection, all of the billions of things that are happening in ads, recommendations, I haven't really even mentioned search yet... Honestly, if I'm right in assuming that the cited figure is really "# of DNNs that do different things", then I'm surprised it's not higher.
Speech recognition barely improved since the 1990s.
We had Dragon natural speaking on a 133MHz Win95 PC (offline of course). After training it for like 10min it worked better or equal as good as Ford's Sync car assistent (offline) and Siri/GoogeNow/Cortana. Well all these services licensed the Nuance speech technology which they got from buying the company behind Dragon natural speaking software. The Ford board computer runs WinCE and has only 233MHz and is still sold in many 2016 Ford cars around the world. And with cloud hosting, to scale the service each users gets only a small amount of total CPU timeslice anyway.
What I want is an offline speech recognition software on my mobile devices! So do I have to install Win95 on an emulator in my smartphone just so my multi-core high end smartphone can do what a Pentium 1 could do in 1996? My hope is on open source projects. Though most such OSS projects are university projects with little documentation how to build the speech model, little community, on an outdated site, written in Java 1.4 and no GitHub page. There is definitely a need for good and competitive C/C++/(native code) TTS and speech recognition project.
[+] [-] peter303|10 years ago|reply
The second boom was expert systems and logic super computers. Those systems never worked that well and A.I. went into long sleep.
Now its supercomputer data mining and greatly improved neural networks.
[+] [-] applecore|10 years ago|reply
[+] [-] mojuba|10 years ago|reply
The academic image recognition machine though seems unstoppable and yes it does seem to improve over time. I honestly don't know what the limits of "dumb" image recognition are in terms of quality, but calling it AI still doesn't make sense to me.
[+] [-] paulsutter|10 years ago|reply
Deepmind's Atari player is based on deep reinforcement learning, where increasing the score represents a reward:
https://m.youtube.com/watch?v=EfGD2qveGdQ
I used to believe the moving goalpost idea, that AI is anything that isn't yet possible. I now disagree.
I saved myself a lot of energy by avoiding entirely the question "what counts as AI?" by switching to the question, "what counts as AGI?" which is a term with a clearer threshold:
https://en.m.wikipedia.org/wiki/Artificial_general_intellige...
[+] [-] gnaritas|10 years ago|reply
That's the core problem of AI, no matter what progress is made, it's instantly called not AI anymore while the goalpost of what AI is is continually being pushed out to not this. The real issue of course is what don't how intelligence actually works so it's impossible to set a fixed goalpost of when AI is truly achieved.
[+] [-] rdtsc|10 years ago|reply
If you don't know how it works -- it looks like magic. It can tell a donkey from a horse, it can play checkers, diagnose a patient etc.
After you are told the trick -- it is just A*, or Rete algorithm, or a multi-layer NN. It makes it less magic and it becomes just another algorithm.
[+] [-] CyberDildonics|10 years ago|reply
[+] [-] Houshalter|10 years ago|reply
The reason it's AI is because it isn't specific to speech recognition. Deep learning is very general. The same algorithms work just as well at speech recognition, or translation, or controlling robots, etc. Image recognition is just the most popular application.
[+] [-] enave|10 years ago|reply
But as soon as chess programs got good, we all took them for granted.
[+] [-] ig1|10 years ago|reply
[+] [-] intrasight|10 years ago|reply
[+] [-] Houshalter|10 years ago|reply
60 years later, we have finally made general purpose learning algorithms, vaguely inspired by the brain, which are just powerful enough to do it. And because they are general purpose, they can also do many other things as well. Everything from speech recognition, to translating sentences, or even controlling robots. Image recognition is just one of many benchmarks that can be used to measure progress.
[+] [-] tim333|10 years ago|reply
[+] [-] such_a_casual|10 years ago|reply
[+] [-] zhanwei|10 years ago|reply
[+] [-] ced|10 years ago|reply
[+] [-] hiddencost|10 years ago|reply
Their machine translation system probably has a similar # of DNNs, and there you have to deal with language pairs, rather than single languages. Let's call it another 400.
That's two side-projects. Then you pull in query prediction, driverless cars, all kinds of infrastructure modeling, spam detection, all of the billions of things that are happening in ads, recommendations, I haven't really even mentioned search yet... Honestly, if I'm right in assuming that the cited figure is really "# of DNNs that do different things", then I'm surprised it's not higher.
[+] [-] n0us|10 years ago|reply
[+] [-] bsder|10 years ago|reply
[+] [-] frik|10 years ago|reply
We had Dragon natural speaking on a 133MHz Win95 PC (offline of course). After training it for like 10min it worked better or equal as good as Ford's Sync car assistent (offline) and Siri/GoogeNow/Cortana. Well all these services licensed the Nuance speech technology which they got from buying the company behind Dragon natural speaking software. The Ford board computer runs WinCE and has only 233MHz and is still sold in many 2016 Ford cars around the world. And with cloud hosting, to scale the service each users gets only a small amount of total CPU timeslice anyway.
What I want is an offline speech recognition software on my mobile devices! So do I have to install Win95 on an emulator in my smartphone just so my multi-core high end smartphone can do what a Pentium 1 could do in 1996? My hope is on open source projects. Though most such OSS projects are university projects with little documentation how to build the speech model, little community, on an outdated site, written in Java 1.4 and no GitHub page. There is definitely a need for good and competitive C/C++/(native code) TTS and speech recognition project.
[+] [-] mrdrozdov|10 years ago|reply
[+] [-] drdeca|10 years ago|reply