Jim, Sorry, we got tangled up in different references.... I was commenting on a link to "kurzweilaI" about "why my phone doesn't understand me" posted by Dr Roberts Jr, but I think you were commenting on a talk by Kurzweil which I have not seen.
Anyway my only point in all this is that while deep learning considered in isolation may not be tantamount to AGI it can still be used in an overall architecture. Everybody realized this, I think, but there is nothing lost in emphasizing it and not get caught in the trap of "either this technique solves AGI *OR* we are not making progress toward AGI." Mike On 1/18/16, Jim Bromer <[email protected]> wrote: > Mike Said: > The statement: > > “Apple’s Siri focuses on statistical regularities, but communication > is not about statistical regularities,” he said. “Statistical > regularities may get you far, but it is not how the brain does it. In > order for computers to communicate with us, they would need a > cognitive architecture that continuously captures and updates the > conceptual space shared with their communication partner during a > conversation.” > > Well, to me that sounds a bit like a false dichotomy. We could use > both (what he calls) statistical regularities working within a > cognitive architecture framework. > > Mike > ------------------------------------------------- > Kurzweil said that computers would need x to communicate with someone > during conversation. He did not say that an assessment of statistical > regularities would need to be excluded to achieve this, so I did not > see his statement as a dichotomy other than to say that statistical > regularities were not enough. > > When I first started listening to the Searle Google Talk I thought > that he was not going to be so iconoclastic about his one main issue. > I agree with him about a lot of the things he said. Computers are > syntactic, we do not understand how consciousness works. His comments > did help me to rethink my plans a little in a way that might lead to > more practical results. However, I realized that we have a fundamental > difference because he thinks that the human brain is not a syntactic > device. Furthermore using Searle's dichotomy we can say that human > consciousness is (subjective) observer relative. In spite of the fact > that we do not know how the mind produces consciousness and regardless > of the mysteriousness of conscious experience, our experience of > consciousness is still relative to our observation (just as our > feelings that a computer is doing thinking when it does some > computation is relative.) > > I don't want to spend the time to get more precise quotes from the > Searle talk, but, using my recall Searle started by talking about the > dichotomy of epistemological knowledge and ontological knowledge. > Ontological knowledge is knowledge that comes from existence and > epistemological knowledge is more like knowledge that has been written > down or derived from higher abstractions. And he says, even though he > realizes that computers can do amazing things, that the computer is > syntactic but it does not have any semantic knowledge about anything. > I do not agree that Searle's dichotomies are absolute. Syntactic > knowledge does contain some semantic information and we can represent > semantic knowledge using syntax. Epistemological knowledge can be used > to encode ontological knowledge. And the brain is a syntactic device. > > So I think that Searle's dichotomies are excessive even though I agree > with some of his view points and I feel that they can be used to help > produce more effective results. In contrast, I don't see a dichotomy > in what Kurzweil was saying unless you are saying that the observation > and utilization of statistical regularities is enough to produce a > cognitive architecture capable of true conversation between computers > and people. > Jim Bromer > > > On Mon, Jan 18, 2016 at 7:00 PM, Mike Archbold <[email protected]> wrote: >> On 1/17/16, Jim Bromer <[email protected]> wrote: >>> The article by Kurzweil seemed to be insightful. >>> >> >> >> To me it sounded like another take on the combinatorial explosion >> issue, which is well known, coming from the angle of the observed >> neural structure of context. >> >> The statement: >> >> “Apple’s Siri focuses on statistical regularities, but communication >> is not about statistical regularities,” he said. “Statistical >> regularities may get you far, but it is not how the brain does it. In >> order for computers to communicate with us, they would need a >> cognitive architecture that continuously captures and updates the >> conceptual space shared with their communication partner during a >> conversation.” >> >> Well, to me that sounds a bit like a false dichotomy. We could use >> both (what he calls) statistical regularities working within a >> cognitive architecture framework. >> >> Mike >> >> >> >> >>> Jim Bromer >>> >>> On Thu, Jan 14, 2016 at 3:24 PM, Raymond D Roberts Jr. via AGI < >>> [email protected]> wrote: >>> >>>> >>>> http://www.kurzweilai.net/why-doesnt-my-phone-understand-me-yet?utm_source=KurzweilAI+Daily+Newsletter&utm_campaign=0481d44bf4-UA-946742-1&utm_medium=email&utm_term=0_6de721fb33-0481d44bf4-282058098 >>>> >>>> Raymond D. Roberts Jr. >>>> >>>> >>>> -----Original Message----- >>>> From: Jim Bromer <[email protected]> >>>> To: AGI <[email protected]> >>>> Sent: Thu, Jan 14, 2016 3:16 pm >>>> Subject: Re: [agi] Re: If Deep Learning is It then Why Are Search >>>> Engines >>>> Incapable of Thinking (Outside the Box or Otherwise)? >>>> >>>> I don't think that deep learning only applies to pure deep learning. It >>>> could be used as part of a system which is attuned to discovering >>>> relations. And it could return relationships (in language for example) >>>> which could then be evaluated. Supervised learning is part of machine >>>> and >>>> deep learning so a system which returns candidate samples that can be >>>> evaluated still could be classified as machine learning (or could be >>>> said >>>> to have something in common with machine learning). >>>> >>>> Older AI paradigms have a much more fixed definition than newer ones. >>>> Comparing Watson, which was apparently able to learn new things about >>>> language, to an old Expert System does not sound right. My argument is >>>> that >>>> almost all contemporary AI paradigms involve some kind of network of >>>> relations so to presuppose that an advanced nlp program cannot 'learn' >>>> about nlp does not make sense. When I get some time I will ask someone >>>> at >>>> IBM what the phrase "Deep NLP" denotes. Does it mean deep search nlp or >>>> does it mean something that is closer to deep learning nlp because >>>> there >>>> is >>>> no reason to rule the possibility that new relationships in nlp could >>>> be >>>> detected in an applied network (of some kind) and then be used as an >>>> abstraction to search for other cases that might have a similar *kind* >>>> of >>>> relationship. >>>> >>>> I am interested in what you said Ben but I get the sense that Watson >>>> was >>>> used to detect relationships in nlp which were then evaluated (probably >>>> in >>>> different ways.) >>>> >>>> Jim Bromer >>>> >>>> On Thu, Jan 14, 2016 at 10:30 AM, Ben Goertzel <[email protected]> >>>> wrote: >>>> >>>>> *** >>>>> My question was why haven't there been clear advances in search engine >>>>> technology in the 2 years since Deep Learning and Watson have made >>>>> very >>>>> obvious advances in AI? >>>>> *** >>>>> >>>>> The Web is very big. Internally within Google, the API calls you can >>>>> make against the whole Web are many fewer than the ones you can make >>>>> against, say, Wikipedia >>>>> >>>>> But business-wise, there is more $$ to be made in making crude >>>>> searches >>>>> against the whole Web slightly less crude, than in making more refined >>>>> and >>>>> intelligent searches against smaller text-bases... >>>>> >>>>> Also, >>>>> >>>>> -- Watson is basically an expert system, albeit a very clever one.... >>>>> Expert system methods don't scale, not even fancy ones... >>>>> >>>>> -- Deep learning in its current form works best for high-dimensional >>>>> floating point data, not discrete data like text .... Also, current >>>>> deep >>>>> learning algorithms rely essentially on bottom-up pattern recognition, >>>>> with >>>>> limited top-down feedback. But real language understanding can't get >>>>> approximated very well without sophisticated top-down feedback.... >>>>> I.e., >>>>> image and speech understanding can get further without cognitive >>>>> feedback, >>>>> than language understanding... >>>>> >>>>> >>>>> ... ben >>>>> >>>>> >>>>> >>>>> On Thu, Jan 14, 2016 at 11:22 PM, Jim Bromer <[email protected]> >>>>> wrote: >>>>> >>>>>> >>>>>> http://www.ibm.com/blogs/think/2016/01/14/the-next-grand-challenge-computers-that-converse-like-people/ >>>>>> >>>>>> Jim Bromer >>>>>> >>>>>> On Thu, Jan 14, 2016 at 10:20 AM, Jim Bromer <[email protected]> >>>>>> wrote: >>>>>> >>>>>>> I watched the Presenti presentation on youtube a few days ago. >>>>>>> >>>>>>> Neural Networks can learn but they cannot use that learning >>>>>>> efficiently >>>>>>> in many important ways. Discrete AI can acquire more specific >>>>>>> (discrete) 'objects' as they learn. So back in the 90's people >>>>>>> started >>>>>>> using hybrids that combined neural networks with discrete methods. >>>>>>> Machine >>>>>>> learning includes advances on hybrid methods. >>>>>>> >>>>>>> Most discrete methods are built around networks of relations between >>>>>>> the data objects which represent 'concepts' or 'ideas' or >>>>>>> 'knowledge' >>>>>>> or >>>>>>> 'know how' or whatever it is that you want to call the data objects >>>>>>> that >>>>>>> would be used to hold knowledge in a (more) discrete AI program. So >>>>>>> a >>>>>>> contemporary discrete AI program is also going to be an >>>>>>> implementation >>>>>>> of a >>>>>>> network. The network may include numerical values but even if it >>>>>>> doesn't >>>>>>> it probably will represent categories of association. That >>>>>>> definition >>>>>>> is >>>>>>> not meant to be complete because I am only trying to get an idea >>>>>>> across: >>>>>>> Modern discrete AI methods involve network methods that can >>>>>>> potentially >>>>>>> be >>>>>>> seen as representatives of 'thought' that are more sophisticated >>>>>>> than >>>>>>> neural networks. That makes sense. >>>>>>> >>>>>>> Presenti was talking about an IBM researcher from the 70s who found >>>>>>> that he could use statistical methods to *learn* about speech >>>>>>> without >>>>>>> a >>>>>>> linguist. That would be a form of machine learning. Therefore it is >>>>>>> fairly >>>>>>> safe for me to conclude that Watson used machine learning in what >>>>>>> Watson >>>>>>> researcher's called, "Deep NLP". >>>>>>> >>>>>>> My question was why haven't there been clear advances in search >>>>>>> engine >>>>>>> technology in the 2 years since Deep Learning and Watson have made >>>>>>> very >>>>>>> obvious advances in AI? I did an image search for "cats" on google >>>>>>> and >>>>>>> it >>>>>>> was very good. I only found one dog (a small dog which had been >>>>>>> photo >>>>>>> shopped with multiple legs somewhat like a caterpillar.) I tried >>>>>>> some >>>>>>> other >>>>>>> searches on images and the results were also very good. The results >>>>>>> were >>>>>>> really amazing. So there have been some advances on image searches >>>>>>> in >>>>>>> the >>>>>>> past 2 years. The search for "castles on the moon" did not >>>>>>> distinguish >>>>>>> between castles pictured as being on the moon from castles with the >>>>>>> moon in >>>>>>> the scene. So even though I am nit-picking to some extent the point >>>>>>> is >>>>>>> that >>>>>>> it looks like you have to train a deep learning neural network with >>>>>>> a >>>>>>> narrow training sample in order to teach it to recognize something >>>>>>> that >>>>>>> would require a little thinking outside the box. That was also a >>>>>>> problem >>>>>>> with Watson. Its Deep NLP could be trained with all the questions >>>>>>> from >>>>>>> past >>>>>>> Jeopardy shows (and Jeopardy-style questions that researchers could >>>>>>> create) >>>>>>> but can it be trained to handle juxtapositions of linguistic >>>>>>> 'concepts' >>>>>>> that might require some thinking outside of the box? (Incidentally I >>>>>>> tried >>>>>>> "cat in a box" and Google did very well. But when I tried "full >>>>>>> stadium" it >>>>>>> did include pictures of stadiums that were not empty. I could spot >>>>>>> them >>>>>>> as >>>>>>> I was paging quickly through the images.) But I guess I there have >>>>>>> been >>>>>>> some significant advances in the past 2 years. They just do not >>>>>>> include >>>>>>> using language to refine your searches. >>>>>>> >>>>>>> My idea of Concept Integration is that different concepts cannot >>>>>>> always >>>>>>> be merged, as in a neural network to take an example, because as >>>>>>> more >>>>>>> concepts are integrated the requirements of a part of the conceptual >>>>>>> integration may change. To restate that in another way,the >>>>>>> integration >>>>>>> of a >>>>>>> number of concepts will typically change if additional concepts are >>>>>>> integrated with them. This is what would happen if you tried to >>>>>>> refine >>>>>>> your >>>>>>> search using conversation. >>>>>>> >>>>>>> Jim Bromer >>>>>>> >>>>>>> >>>>>>> On Tue, Jan 12, 2016 at 10:21 PM, LAU <> wrote: >>>>>>> >>>>>>>> OK, as you wish ... It's just a word. We do not agree on the >>>>>>>> signification. But, it's OK. If you call it "deep learning", or >>>>>>>> "conceptual >>>>>>>> learning" ... or "Hakuna Matata's learning", it's not important. >>>>>>>> Stop >>>>>>>> playing with words. >>>>>>>> >>>>>>>> Try to back to the topic of this thread. If I understand what you >>>>>>>> want >>>>>>>> to promote : >>>>>>>> 1) You note that deep learning implemented in the industry is not >>>>>>>> so >>>>>>>> intelligent than espected taking in account the computation power >>>>>>>> available. >>>>>>>> 2) Watson seems to less "narrow" than other implementations. >>>>>>>> 3) What it miss there is "conceptual integration". >>>>>>>> Correct me if I'm wrong. >>>>>>>> >>>>>>>> >>>>>>>> In my humble opinion, there's no intelligent machines just because >>>>>>>> people don't try to, or most likely don't figure out how to, make >>>>>>>> it >>>>>>>> more >>>>>>>> intelligent. >>>>>>>> Implementing"conceptual integration" is certainly a way that some >>>>>>>> researchers tried, but lead to no significant results until now. If >>>>>>>> I >>>>>>>> look >>>>>>>> at wikipedia, the theory is dated from 1990s. Twenty years later, >>>>>>>> still >>>>>>>> nothing. >>>>>>>> >>>>>>>> There's no magic behind deep learning, I mean neural networks, used >>>>>>>> by >>>>>>>> Google or Facebook. Very roughly, it's just a kind of "universal >>>>>>>> approximator". And it's not the computation power that with make it >>>>>>>> spontaneously more intelligent. >>>>>>>> Deep learning becomes very popular these last years because it's >>>>>>>> easier to make a neural network to accomplish picture or voice >>>>>>>> recognition >>>>>>>> task (*I've made a small one myself from scratch in few days*) than >>>>>>>> handcrafted codes, and for a better result. >>>>>>>> But basically, a neural network is just another kind of >>>>>>>> programming. >>>>>>>> Instead of coding a multitude of operation to achieve a complex >>>>>>>> task, >>>>>>>> a >>>>>>>> neural network can do it itself by learning from examples. >>>>>>>> And the question will be how to teach a neural network what is >>>>>>>> "conceptual integration" ? >>>>>>>> >>>>>>>> In the paris tech conference video (*on youtube, but it's in french >>>>>>>> ...*), Jerome Pesenti said something else interesting. He cite a >>>>>>>> IBM's 70s researcher, Fred Jelinek, who said "*Every time I fire a >>>>>>>> linguist the performance of the speech recognizer goes up*". The >>>>>>>> Jelinek speech recognizer team was composed by part of linguist and >>>>>>>> engineers. By replacing a linguist who treats language as a human >>>>>>>> do >>>>>>>> by an >>>>>>>> engineer who does mathematics and statistics on words, the result >>>>>>>> is >>>>>>>> better. And it seems to be the philosophy at IBM to work >>>>>>>> differently >>>>>>>> that a >>>>>>>> human do, and it seems to give better result. Instead of playing >>>>>>>> jeopardy >>>>>>>> in human way, watson applies statistics on the database (*which was >>>>>>>> wikipedia*). >>>>>>>> >>>>>>>> What I want to say is that may be the "conceptual integration" is a >>>>>>>> track to explore for building AGI. Or, may be the solution will >>>>>>>> come >>>>>>>> from >>>>>>>> elsewhere. >>>>>>>> >>>>>>>> >>>>>>>> LAU >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Le 12/01/2016 10:46, Jim Bromer a écrit : >>>>>>>> >>>>>>>> Deep Learning is Deep Machine Learning and Machine Learning is in >>>>>>>> no >>>>>>>> way limited to Neural Networks. So there is no way that Deep >>>>>>>> Learning >>>>>>>> is going to be forever defined to refer Machine Learning that uses >>>>>>>> Neural Networks (in certain ways). From that point of view I can >>>>>>>> say >>>>>>>> that Watson-Jeopardy probably did use a kind of deep learning. >>>>>>>> >>>>>>>> >>>>>>>> ------------------------------------------- >>>>>>>> AGI >>>>>>>> Archives: https://www.listbox.com/member/archive/303/=now >>>>>>>> RSS Feed: >>>>>>>> https://www.listbox.com/member/archive/rss/303/27172223-36de8e6c >>>>>>>> Modify Your Subscription: https://www.listbox.com/member/?&; >>>>>>>> Powered by Listbox: http://www.listbox.com >>>>>>>> >>>>>>>> >>>>>>>> *AGI* | Archives | Modify Your Subscription >>>>>>>> >>>>>>> >>>>>>> >>>>>> *AGI* | Archives | Modify Your Subscription >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> Ben Goertzel, PhD >>>>> http://goertzel.org >>>>> >>>>> "The reasonable man adapts himself to the world: the unreasonable one >>>>> persists in trying to adapt the world to himself. Therefore all >>>>> progress >>>>> depends on the unreasonable man." -- George Bernard Shaw >>>>> *AGI* | Archives | Modify Your Subscription >>>>> >>>> >>>> *AGI* | Archives | Modify Your Subscription >>>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>>> <https://www.listbox.com/member/archive/rss/303/24379807-653794b5> | >>>> Modify >>>> <https://www.listbox.com/member/?&;> >>>> Your Subscription <http://www.listbox.com> >>>> >>> >>> >>> >>> ------------------------------------------- >>> AGI >>> Archives: https://www.listbox.com/member/archive/303/=now >>> RSS Feed: >>> https://www.listbox.com/member/archive/rss/303/11943661-d9279dae >>> Modify Your Subscription: >>> https://www.listbox.com/member/?&; >>> Powered by Listbox: http://www.listbox.com >>> >> >> >> ------------------------------------------- >> AGI >> Archives: https://www.listbox.com/member/archive/303/=now >> RSS Feed: >> https://www.listbox.com/member/archive/rss/303/24379807-653794b5 >> Modify Your Subscription: https://www.listbox.com/member/?&; >> Powered by Listbox: http://www.listbox.com > > > ------------------------------------------- > AGI > Archives: https://www.listbox.com/member/archive/303/=now > RSS Feed: https://www.listbox.com/member/archive/rss/303/11943661-d9279dae > Modify Your Subscription: > https://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
