You are viewing a single comment's thread from:

RE: The value of Knowledge Representation and the Decentralized Knowledge Base for Artificial Intelligence (expert systems)

in #ai • 7 years ago

Hmm, as a person with no programming experience and who was just watching a video of a kitten riding a turtle, this piece was a little demanding on my pea sized brain

In addition to greater efficiency in building a knowledge base, I believe one of the greatest advantages of the decentralization of AI is its ability to reduce the probability of minority bad actors successfully harnessing the technology for large scale nefarious ends.

Also, I did not follow your reasoning on Part 4. The paper you cited concluded that at times there is a trade off between expressiveness and computational efficiency, yet you interpreted this as 'better representational knowledge leads to computational efficiency.' Surely in certain cases, for example telling a joke, highly efficient methods of representation such as syllogistic logic is not the best mode of delivery? If chicken crosses road Then chicken wanted to get to the other side.

Very interesting read as usual

7 years ago in #ai by
$1.11
  • Past Payouts $1.11, 0.00 TRX
  • - Author $1.08, 0.00 TRX
  • - Curators $0.03, 0.00 TRX
4 votes
Reply 2
Sort:  
  • Trending
    • [-]
     7 years ago (edited)

    Yes that particular quote is a bit cryptic and not easy for me to explain. You always have trade offs and yes expressiveness can cost you efficiency.

    This quote is perhaps much better:

    One contribution to this pragmatic efficiency is supplied by the guidance a representation provides for organizing information so as to facilitate making the recommended inferences.

    Which I interpreted (and this interpretation could be wrong) that KR contributes to pragmatic efficiency through how it's organized such that it can facilitate RECOMMENDED INFERENCES.

    And then we have this quote below:

    While sanctioned inferences tell us what conclusions we are permitted to make, that set is invariably very large and hence provides insufficient constraint. Any automated system attempting to reason, guided only by knowing what inferences are sanctioned, will soon find itself overwhelmed by choices. Hence we need more than an indication of which inferences we can legally make, we also need some indication of which inferences which inferences are appropriate to make, i.e., intelligent. That indication is supplied by the set of recommended inferences.

    So in other words I think by efficiency they are talking about reducing or narrowing down the inferences from that which is sanctioned (available to be chosen) to that which is intelligent and or appropriate to choose. I guess I would think of it as common sense?

    This last quote is also vital:

    Note that the need for a specification of recommended inferences means that in specifying a representation we also need to say something about how to reason intelligently. Representation and reasoning are inextricably, and usefully, intertwined: a knowledge representation is a theory of intelligent reasoning.

    Humans are good at common sense and can do humor. Machines have a very difficult time with that because by default a machine would not narrow the search space or know certain stereotypes or common human inferences for certain situations. It's the difference between sanctioned and recommended, between reason and common sense.

    References

    1. http://groups.csail.mit.edu/medg/ftp/psz/k-rep.html
    $0.30
    • Past Payouts $0.30, 0.00 TRX
    • - Author $0.27, 0.00 TRX
    • - Curators $0.03, 0.00 TRX
    1 vote
    Reply
    [-]
     7 years ago 

    😍😵😂

    Finally a bitta craic ;)

    Very excited for what you bring to steemit ❤

    $0.06
    • Past Payouts $0.06, 0.00 TRX
    • - Author $0.05, 0.00 TRX
    • - Curators $0.01, 0.00 TRX
    2 votes
    Reply

    Coin Marketplace

    STEEM 0.16
    TRX 0.15
    JST 0.028
    BTC 56160.19
    ETH 2367.48
    USDT 1.00
    SBD 2.31