In the 60s the dominant ideology was that symbolic manipulation was all there is to AI (Society of Mind)[1]. Lisp (List Processing language) was the obvious choice for this as it is very good for string manipulation (emacs and elisp are proof of this). Also the ability to express code as lists or s-expressions made it possible to write program-writing-programs or macros which turned out to be very effective for text crunching.
Today, statistical learning is considered to be a better approach to building an AI. Much of the techniques of ML are expressed easily and efficiently using linear algebra for which Matlab (Matrix Labaratory) or Octave is much better suited.
I think Lisp fell out of favor as it is terrible when it comes to number crunching.
> Today, statistical learning is considered to be a better approach to building an AI. Much of the techniques of ML are expressed easily and efficiently using linear algebra for which Matlab (Matrix Labaratory) or Octave is much better suited.
The first statement is true: statistical learning is considered to be the leading edge of current AI research, especially since it attacks combinatorial explosion head-on. However, linear algebra is the baseline of current work, rarely the driving force. While the 'pull eigenvectors out of shitloads of data" method has been very effective and profitable, it is a very shallow approach to AI (especially considering the problems attacked in the 60s). Luckily, more intelligent algorithms are being developed that combine symbolic and probabilistic reasoning in a more direct way: compare LSA to LDA for example; LDA has a more cohesive and explanatory structure. These kind of things demand flexible programming environments (such as Lisp): Matlab is where algorithms go once they've been proven useful for 'normal people;' it is the domain of the analyst who cares about the data instead of the algorithm itself :).
AI itself fell out of flavor because of social forces, see the AI winter. I don't believe AI has really recovered socially or economically because even now there is a major financial crisis.
bfrs|14 years ago
Today, statistical learning is considered to be a better approach to building an AI. Much of the techniques of ML are expressed easily and efficiently using linear algebra for which Matlab (Matrix Labaratory) or Octave is much better suited.
I think Lisp fell out of favor as it is terrible when it comes to number crunching.
[1] The dramatic collapse of this ideology is chronicled in this wired article (Two AI Pioneers. Two Bizarre Suicides. What Really Happened?): http://www.wired.com/techbiz/people/magazine/16-02/ff_aimyst...
wickedchicken|14 years ago
The first statement is true: statistical learning is considered to be the leading edge of current AI research, especially since it attacks combinatorial explosion head-on. However, linear algebra is the baseline of current work, rarely the driving force. While the 'pull eigenvectors out of shitloads of data" method has been very effective and profitable, it is a very shallow approach to AI (especially considering the problems attacked in the 60s). Luckily, more intelligent algorithms are being developed that combine symbolic and probabilistic reasoning in a more direct way: compare LSA to LDA for example; LDA has a more cohesive and explanatory structure. These kind of things demand flexible programming environments (such as Lisp): Matlab is where algorithms go once they've been proven useful for 'normal people;' it is the domain of the analyst who cares about the data instead of the algorithm itself :).
jhuni|14 years ago
dudeguy999|14 years ago
treetrouble|14 years ago