Re: Feature Engineeriing!
> The mistake is the same in both cases. The solution is not better
> features,
> but to abandon the assumption those features organize themselves in only
> one
> way.
>
> -Rob
Your approach is very interesting Rob, and the one about
retaining many meanings in the natural language representation
rings a strong bell with me. I wish I knew how to take the next
step and find a representation that preserves the many levels
of interpretation that people place on a phrase.
Thanks,
Rich
> On Thursday 03 March 2005 17:59, John F. Sowa wrote:
>> For years, the people who were working with neural nets
>> and other learning systems were hyping the claim that
>> they did away with "knowledge engineers" because they
>> didn't need rules, frames, formulas, etc.
>>
>> But now they discovered that these learning systems don't
>> really learn all by themselves -- somebody has to analyze
>> the subject domain and select some set of "features" for
>> the learning systems to work with. Since they had already
>> got rid of the knowledge engineers, they now discovered
>> a totally new breed of "feature engineers".
>>
>> Come drink the feature soup -- formerly known as knowledge soup.
>>
>> John Sowa
>