4 ms·

The idea that knowledge can be expressed in distributed representations rather than sequences of symbols is a huge advancement in our understanding of intellige

by dlkf 3y ago

The idea that knowledge can be expressed in distributed representations rather than sequences of symbols is a huge advancement in our understanding of intelligence. Obviously we haven’t ”solved intelligence” but now we at least have some idea of what the right questions to ask are.

disgruntledphd2 3y ago
Can you share a link to the paper please? I feel like I get you, but would like to disambiguate with the mathematical representation.
- dlkf 3y ago
  I’m not talking about a single paper, I’m talking about a whole research programme. But Rumelhart, Hinton, Williams 1986 is very improtant. A lot of the foundational work is collected in the PDP report.
  - disgruntledphd2 3y ago
    Links would be greatly appreciated. I'm definitely not familiar with NN work from the 80s.
  - disgruntledphd2 3y ago
    > But Rumelhart, Hinton, Williams 1986 is very improtant. A lot of the foundational work is collected in the PDP report. Thanks! OK, so this is basically a connectionist model of mind approach? I can definitely see this as an ancestor of current neural network approaches, and I now have some idea of what you mean by distributed representations. I mean no disrespect here, but this has not contributed to our understanding of intelligence at all. It's proved useful in getting large datasets to perform certain actions (like vision and speech), but those are not necessarily the same thing at all. It's a massive, massive advancement in the field of statistics and learning from data, but doesn't seem to map to my conception of intelligence at all. (as you may have guessed, I'm sceptical that statistical learning approaches will lead to human-level intelligence).
    - dlkf 3y ago
      > It's a massive, massive advancement in the field of statistics and learning from data, but doesn't seem to map to my conception of intelligence at all. You think ”learning from data” has nothing to do with ”intelligence” ?

by dlkf 3y ago

disgruntledphd2 3y ago

Can you share a link to the paper please? I feel like I get you, but would like to disambiguate with the mathematical representation.

dlkf 3y ago
I’m not talking about a single paper, I’m talking about a whole research programme. But Rumelhart, Hinton, Williams 1986 is very improtant. A lot of the foundational work is collected in the PDP report.
- disgruntledphd2 3y ago
  Links would be greatly appreciated. I'm definitely not familiar with NN work from the 80s.
- disgruntledphd2 3y ago
  > But Rumelhart, Hinton, Williams 1986 is very improtant. A lot of the foundational work is collected in the PDP report. Thanks! OK, so this is basically a connectionist model of mind approach? I can definitely see this as an ancestor of current neural network approaches, and I now have some idea of what you mean by distributed representations. I mean no disrespect here, but this has not contributed to our understanding of intelligence at all. It's proved useful in getting large datasets to perform certain actions (like vision and speech), but those are not necessarily the same thing at all. It's a massive, massive advancement in the field of statistics and learning from data, but doesn't seem to map to my conception of intelligence at all. (as you may have guessed, I'm sceptical that statistical learning approaches will lead to human-level intelligence).
  - dlkf 3y ago
    > It's a massive, massive advancement in the field of statistics and learning from data, but doesn't seem to map to my conception of intelligence at all. You think ”learning from data” has nothing to do with ”intelligence” ?