6 ms·

The fascinating thing about his answers is that he says nothing that wasn't already known 25 years ago. 'General AI' is a really vague term and means almost not

by dpandey 10y ago

The fascinating thing about his answers is that he says nothing that wasn't already known 25 years ago. 'General AI' is a really vague term and means almost nothing. It's well known in AI that common sense AI systems are incredibly hard to build and expert AI systems are relatively straight forward to model and build. Given that the most 'common sense' type products we see like Siri and Alexa still appear to be rules based systems (that have been enhanced with historical data no doubt), we all know the world is not suddenly going to become a common sense AI paradise unless a new Marvin Minsky makes some breakthroughs. I find that the term 'AI' is frequently used when a company has taken an existing problem and enhanced the solution by using machine learning at one or more places. Nothing wrong with that fundamentally, but calling that AI is more of a marketing gimmick than anything else. Google/Facebook do some of the best machine learning enhanced workflows that delight users. They usually refrain from branding it as AI.

_pmf_ 10y ago
What we currently observe is not a revival of AI research, but an emergent mutual exponentiation of large-scale data collection and classic statistical AI techniques. They now have a nearly infinitely customizable source of training data and verification data that can be used to train bog standard neural nets.
- dpandey 10y ago
  Agree, except with the 'nearly infinite' qualifier. We have some large data sets now and we can draw a cat. We don't know if we can build anything interesting with those data sets yet.
- rspeer 10y ago
  This is essentially true, and I'm sad that someone is so caught up in the hype that they'd downvote a simple description of the actual state of the field. What we have right now is a boom in AI engineering, with AI research fairly neglected because it doesn't produce the same kind of hype. Here's a Twitter thread [1] where NLP researchers describe people who work on machine translation as 'depressed', because they spend their time tweaking neural net parameters to get tiny improvements over the best techniques of last decade, with no opportunity to understand what causes an improvement. [1] https://twitter.com/haldaume3/status/805055952789573635 https://twitter.com/haldaume3/status/805055952789573635
randcraw 10y ago
I think the term GAI/AGI isn't meaningless. It's just imprecise. But the phrase has value since there is no better term for "human-level cognitive ability". I'm pretty sure Siri/Alexa/GoogleNow don't use rules. Like all modern Q&A systems like IBM Watson, they use a variety of statistical word match/associations and phrase/sentence response templates. Like Watson, the algorithm identifies the most probable word match to fill the void in your query and then inserts it into the most appropriate answer template. Yes, it's a bit hit or miss (like Watson answering "Toronto" for the US city with an airport named after a WWII hero). But if Apple/Google/Amazon adds some simple sequential disambiguation and refinement (clarification Q&A), I suspect this approach should answer most casual questions quite nicely. IMHO, calling it "AI" means that the problem is not amenable to a closed form solution, so some form of iterative search and refinement is required. In venerable GOFAI parlance, AI problems are inherently 'scruffy'. Once the problem can be solved efficiently, it becomes 'neat' and it's no longer considered to be AI. IMHO, most of today's progress in AI is scruffy (probabilistic word matches and deep learning), since they're driven not by formal internally-curated information but by smart search against presumed-relevant external referents, requiring secondary refinement -- a process that is largely inefficient and open-ended and error-prone, and will remain so indefinitely, I think, since we're not trying to improve the quality of the referent data, which is the weakest link in the cognitive chain.
- dpandey 10y ago
  I was just thinking about that. Once you can solve a problem in a way that's currently comprehensible and doesn't feel like 'magic', it's no longer AI. Having said that, I still see abuse of the phrase (more in industry than academia) where standard machine learning applied to a workflow is marketed as some big AI in a product. But I understand why they do it - it usually helps in the long drawn out enterprise sales process to be perceived as an AI powered product.
  - ktta 10y ago
    >I was just thinking about that. Once you can solve a problem in a way that's currently comprehensible and doesn't feel like 'magic', it's no longer AI. https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect
StavrosK 10y ago
Semi-offtopic, but I've always wondered why something like a medical diagnosis AI doesn't exist. The doctor (or operator) would enter all your symptoms and test results, and the software would use statistical data and research results to give you a probability of each disease (e.g. a headache is overwhelmingly probably a flu, but if your test results for X are elevated, the probability for brain cancer increases, etc), as well as a list of tests the doctor could run. Something like this would be immensely valuable, and would immediately improve healthcare everywhere, since the software could be updated with new or obscure knowledge right away, so you wouldn't have to shop around for doctors until you came across someone who had read of a very rare disease somewhere, the software would already know everything you might have. Is there something like that out there? It sounds like there must be, but I don't know of any doctors using anything like this.
- ska 10y ago
  These systems have been worked on since the 90s at least. The latest branding of them is "Decision Support Systems" but there are lots of CAD (Computer Aided Diagnostic, cf Computer Aided Detection). There is a lot of interesting work that falls under the general heading of "Evidence based medicine" that relates to this. There are a lot of barriers to presenting it as you imagine, though: some liability, some workflow, and some just the deep lack of appropriate datasets. There is also some adoption issues - clinicians are rarely going to be happy with a black box approach but it can be difficult to explain/expose the details of such analysis in a way that clinicians will accept. A last thought: this is actually very hard to do well and dangerous to make mistakes with - so your rather cavalier claim that there would be immediate and wide spread improvement of health care should be tempered a bit.
  - dpandey 10y ago
    Your last thought is key. Medical diagnosis demands extremely high precision as well as 'judgement' because one little deviation can be highly correlated with something serious. And there can be several of these deviations in different places. Sometimes a doctor notices something is off while looking at something else. It's not easy to model out all these while keeping up with the changing knowledge in the field and still avoiding false negatives.
deong 10y ago
Yes, the stuff he's saying has been known for ages, but there's been a greater trend recently of people hinting otherwise. I worked in a research group that was focused on AGI until I left academia last year, and there are spirited debates going on about just how close we are. For what it's worth, I completely agree with the sentiment presented here that we're far enough away that we don't know what direction even we need to start down to make the best progress, which in my view makes any sort of prognostication of questionable value. It's also why I think it's not that productive to have people like Stephen Hawking arguing about future AI as a threat that somehow needs to be mitigated. Even if he's right in the general point, we know so little about how that future system would look that it would be like asking Isaac Newton to make cars safer.
- dpandey 10y ago
  Note that he didn't say anything about the point you made : a possible change in common sense systems being more tractable to build. That'd have been fascinating. He just said something old and techcrunch made it into a headline. Love your point about hawking. I treat that as pure entertainment but since the general population treats hawking with high regard (well deserved) his opinions about things he doesn't know much about also carry weight in shaping public perception. As far as speculating on how scary an AI driven future can be, a filmmaker in the 60s or a high school student today can do as compelling a job as Hawking.
- YeGoblynQueenne 10y ago
  >> we're far enough away that we don't know what direction even we need to start down to make the best progress Even worse, we have some strong theoretical results suggesting that none of the techniques we have developed so far can ever lead to AGI, and that refers very much to machine learning primarily. I'm talking about the work that was done on inductive inference, and then PAC-learning, in decades past (starting in the 1960's). There are many negative results and very few positive ones. I wouldn't say I've reviewed the bibliography extensively (it's a huge area spanning a couple of fields), but from the little I've read I'm getting the feeling that, if you want to be able to learn any function in polynomial time, with a polynomial number of training examples, and with arbitrary error, you're basically stuck with the easiest classes of concept to learn. And if you want to learn specific distributions or representations (like DFAs or PCFGs) then you can just forget about it. Which reminds me of something that Andrew Ng has said, about how recent results in machine learning and deep learning only reached the "low hanging fruit" and further progress is going to be much harder. No references to all this- I'll dig 'em up if required.

by dpandey 10y ago

_pmf_ 10y ago

What we currently observe is not a revival of AI research, but an emergent mutual exponentiation of large-scale data collection and classic statistical AI techniques. They now have a nearly infinitely customizable source of training data and verification data that can be used to train bog standard neural nets.

dpandey 10y ago
Agree, except with the 'nearly infinite' qualifier. We have some large data sets now and we can draw a cat. We don't know if we can build anything interesting with those data sets yet.
rspeer 10y ago
This is essentially true, and I'm sad that someone is so caught up in the hype that they'd downvote a simple description of the actual state of the field. What we have right now is a boom in AI engineering, with AI research fairly neglected because it doesn't produce the same kind of hype. Here's a Twitter thread [1] where NLP researchers describe people who work on machine translation as 'depressed', because they spend their time tweaking neural net parameters to get tiny improvements over the best techniques of last decade, with no opportunity to understand what causes an improvement. [1] https://twitter.com/haldaume3/status/805055952789573635 https://twitter.com/haldaume3/status/805055952789573635

randcraw 10y ago

I think the term GAI/AGI isn't meaningless. It's just imprecise. But the phrase has value since there is no better term for "human-level cognitive ability". I'm pretty sure Siri/Alexa/GoogleNow don't use rules. Like all modern Q&A systems like IBM Watson, they use a variety of statistical word match/associations and phrase/sentence response templates. Like Watson, the algorithm identifies the most probable word match to fill the void in your query and then inserts it into the most appropriate answer template. Yes, it's a bit hit or miss (like Watson answering "Toronto" for the US city with an airport named after a WWII hero). But if Apple/Google/Amazon adds some simple sequential disambiguation and refinement (clarification Q&A), I suspect this approach should answer most casual questions quite nicely. IMHO, calling it "AI" means that the problem is not amenable to a closed form solution, so some form of iterative search and refinement is required. In venerable GOFAI parlance, AI problems are inherently 'scruffy'. Once the problem can be solved efficiently, it becomes 'neat' and it's no longer considered to be AI. IMHO, most of today's progress in AI is scruffy (probabilistic word matches and deep learning), since they're driven not by formal internally-curated information but by smart search against presumed-relevant external referents, requiring secondary refinement -- a process that is largely inefficient and open-ended and error-prone, and will remain so indefinitely, I think, since we're not trying to improve the quality of the referent data, which is the weakest link in the cognitive chain.

dpandey 10y ago
I was just thinking about that. Once you can solve a problem in a way that's currently comprehensible and doesn't feel like 'magic', it's no longer AI. Having said that, I still see abuse of the phrase (more in industry than academia) where standard machine learning applied to a workflow is marketed as some big AI in a product. But I understand why they do it - it usually helps in the long drawn out enterprise sales process to be perceived as an AI powered product.
- ktta 10y ago
  >I was just thinking about that. Once you can solve a problem in a way that's currently comprehensible and doesn't feel like 'magic', it's no longer AI. https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect

StavrosK 10y ago

Semi-offtopic, but I've always wondered why something like a medical diagnosis AI doesn't exist. The doctor (or operator) would enter all your symptoms and test results, and the software would use statistical data and research results to give you a probability of each disease (e.g. a headache is overwhelmingly probably a flu, but if your test results for X are elevated, the probability for brain cancer increases, etc), as well as a list of tests the doctor could run. Something like this would be immensely valuable, and would immediately improve healthcare everywhere, since the software could be updated with new or obscure knowledge right away, so you wouldn't have to shop around for doctors until you came across someone who had read of a very rare disease somewhere, the software would already know everything you might have. Is there something like that out there? It sounds like there must be, but I don't know of any doctors using anything like this.

ska 10y ago
These systems have been worked on since the 90s at least. The latest branding of them is "Decision Support Systems" but there are lots of CAD (Computer Aided Diagnostic, cf Computer Aided Detection). There is a lot of interesting work that falls under the general heading of "Evidence based medicine" that relates to this. There are a lot of barriers to presenting it as you imagine, though: some liability, some workflow, and some just the deep lack of appropriate datasets. There is also some adoption issues - clinicians are rarely going to be happy with a black box approach but it can be difficult to explain/expose the details of such analysis in a way that clinicians will accept. A last thought: this is actually very hard to do well and dangerous to make mistakes with - so your rather cavalier claim that there would be immediate and wide spread improvement of health care should be tempered a bit.
- dpandey 10y ago
  Your last thought is key. Medical diagnosis demands extremely high precision as well as 'judgement' because one little deviation can be highly correlated with something serious. And there can be several of these deviations in different places. Sometimes a doctor notices something is off while looking at something else. It's not easy to model out all these while keeping up with the changing knowledge in the field and still avoiding false negatives.

deong 10y ago

Yes, the stuff he's saying has been known for ages, but there's been a greater trend recently of people hinting otherwise. I worked in a research group that was focused on AGI until I left academia last year, and there are spirited debates going on about just how close we are. For what it's worth, I completely agree with the sentiment presented here that we're far enough away that we don't know what direction even we need to start down to make the best progress, which in my view makes any sort of prognostication of questionable value. It's also why I think it's not that productive to have people like Stephen Hawking arguing about future AI as a threat that somehow needs to be mitigated. Even if he's right in the general point, we know so little about how that future system would look that it would be like asking Isaac Newton to make cars safer.

dpandey 10y ago
Note that he didn't say anything about the point you made : a possible change in common sense systems being more tractable to build. That'd have been fascinating. He just said something old and techcrunch made it into a headline. Love your point about hawking. I treat that as pure entertainment but since the general population treats hawking with high regard (well deserved) his opinions about things he doesn't know much about also carry weight in shaping public perception. As far as speculating on how scary an AI driven future can be, a filmmaker in the 60s or a high school student today can do as compelling a job as Hawking.
YeGoblynQueenne 10y ago
>> we're far enough away that we don't know what direction even we need to start down to make the best progress Even worse, we have some strong theoretical results suggesting that none of the techniques we have developed so far can ever lead to AGI, and that refers very much to machine learning primarily. I'm talking about the work that was done on inductive inference, and then PAC-learning, in decades past (starting in the 1960's). There are many negative results and very few positive ones. I wouldn't say I've reviewed the bibliography extensively (it's a huge area spanning a couple of fields), but from the little I've read I'm getting the feeling that, if you want to be able to learn any function in polynomial time, with a polynomial number of training examples, and with arbitrary error, you're basically stuck with the easiest classes of concept to learn. And if you want to learn specific distributions or representations (like DFAs or PCFGs) then you can just forget about it. Which reminds me of something that Andrew Ng has said, about how recent results in machine learning and deep learning only reached the "low hanging fruit" and further progress is going to be much harder. No references to all this- I'll dig 'em up if required.