On Fri, Jan 4, 2013 at 10:34 PM, Stan Nilsen <[email protected]> wrote:
> > On 01/04/2013 05:42 PM, Anastasios Tsiolakidis wrote: > >> On Fri, Jan 4, 2013 at 10:25 PM, Piaget Modeler >> <[email protected] >> <mailto:piagetmodeler@hotmail.**com<[email protected]>>> >> wrote: >> >> >> http://www.marcpickett.com/**RepLearn2013/<http://www.marcpickett.com/RepLearn2013/> >> >> >> Right, they prefer automatically derived rather than hand-crafted >> "learning". I just want to be on record again saying "it can't be done". >> Sure, General Intelligence excels at finding structure in unstructured >> data and enriching "fluid", continuous domains with rather quantized >> concepts and superstructures. But for probabilistic and other reasons I >> do not expect a blank slate to develop anthropomorphic concepts, ever. >> >> AT >> > > I tend to agree that "automatically derived" is not the way to go. The > knowledge is needed, but we ought to think in terms of "accumulated" rather > than derived. > > ... on a similar note... > > I was pleased to see the way the organizer's phrased the problem - > > "A human-level artificially intelligent agent must be able to represent > and reason about the world, at some level, in terms of high-level concepts > such as entities and relations. The problem of acquiring these rich > high-level representations, known as the knowledge acquisition bottleneck, > has long been an obstacle for achieving human-level AI. " > > It fits well with my recent thoughts about "what is so hard..." In my > thinking, the "knowledge acquisition bottleneck" is the battle to fight. > I refer to the problem as the "Scheme of things" issue. > Most knowledge is stored in human sentences. > > Before an intelligent unit is able to make choices about general things, > the unit will need to see the overall scheme of things. If there is an > option in consideration, how does it "fit" the scheme of things. > > (My opinion )"Scheme of things" knowledge is not being accumulated, in > part, because we don't have a format for the data. We don't have the > format for the data because we don't have a "design" that says "this is how > the data will be used by components x, y, z." My emphasis is on > determining what I think xyz should be. > > Stan > Sentences have been expressed for thousands of years, without necessarily knowing about future people that may utilize them. Yet they still work. > > > > > ------------------------------**------------- > AGI > Archives: > https://www.listbox.com/**member/archive/303/=now<https://www.listbox.com/member/archive/303/=now> > RSS Feed: https://www.listbox.com/**member/archive/rss/303/** > 5037279-a88c7a6d<https://www.listbox.com/member/archive/rss/303/5037279-a88c7a6d> > Modify Your Subscription: https://www.listbox.com/** > member/?&id_**secret=5037279-8beb0005<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
