"Since any neural net needs to be able to interpert feedback as a success, failure, or somthing in between in order to 'learn'"
'Success' and 'failure' are human constructs. To a neural network, they're just connection strengths. The idea here would be simply to copy patterns of connection strengths wholesale from a human brain into a neural net.
In fact, our basic emotional processing wetware was copied into our brains (giant neural nets) via the DNA that we got from our respective parents.
Actually, in his book, from whose title this conference draws its name, Kurzweil does not claim that machines will have spiritual experiences; merely, that machines will claim to have spiritual experiences, and that people will believe them.
My point is that Christianity presents an earth-centric, anthropocentric view of the universe; that its inventors made no provisions for the existence of extraterrestrial intelligence; and that the confirmed existence of ET's will undermine seriously the already eroded credibility of Christianity as a way of knowing about the nature of reality.
After all, intelligent aliens are not animals, nor are they men. They certainly are not descended from Adam and Eve. Genesis, alone, precludes the very existence of such persons.
In the past, the church has modified its doctrine in order to accomodate contrary, irrefutable new knowledge about the nature of reality, such as the actual shape of the earth (though not, in this particular case, before murdering a number of heroically intransigent scientists). Some would no doubt aver that this points to a certain 'flexibility' in churh thinking that should not escape admiration; in my own view, however, it demonstrates that church doctrine is so subject to interpretation as to have no meaning, whatever.
That's my point. Of course, YMMV: that's why they print menus:)
If the economy continues to perform well, I think that many American's may regain the taste for space exploration that they lost during the recession of the 70's.
The kind of session tracking that you get with cookies can also be achieved with URL rewriting. In fact, back in the good old pre-cookie days (before AOL opened its Internet gateway ), this was the only way to do persistence.
Smart servlet engines can detect when a user is refusing cookies, and do URL rewriting, instead. (don't know whether asp, cold fusion, &c do this).
How about a browser feature that refuses all cookies, and autosends a message to the originating site admin telling her to switch to jakarta-tomcat;)
This touches on the very important point that knowing a lot about a subject is very different from having mastery of a subject.
A person who can visualize regions of electron density the way most people can visuzlize legos has achieved mastery. A person who can merely rattle off constants has not.
In order to attain mastery of a subject, one must struggle through lots and lots of details, and undergraduate exams are mostly about testing a student's grasp of these details. But the ultimate goal of advanced education is to produce masters.
I would even go so far as to say that most undergrads needn't worry overly about committing all of these details to memory because they do not yet know which of them they will actually use in their careers (graduate or otherwise). When they start doing real work, they will naturally commit certain oft-used facts to memory. Others, they will just as naturally forget, and never miss them.
What matters in undergraduate education (other than beer;) is that students gain sufficient background knowledge to be able to understand complex problems, and begin to think about them in relevant ways. From this all things proceed.
Slashdot Mirror
'Success' and 'failure' are human constructs. To a neural network, they're just connection strengths. The idea here would be simply to copy patterns of connection strengths wholesale from a human brain into a neural net.
In fact, our basic emotional processing wetware was copied into our brains (giant neural nets) via the DNA that we got from our respective parents.
Actually, in his book, from whose title this conference draws its name, Kurzweil does not claim that machines will have spiritual experiences; merely, that machines will claim to have spiritual experiences, and that people will believe them.
It's the exceptions handling that takes so long to code.
My point is that Christianity presents an earth-centric, anthropocentric view of the universe; that its inventors made no provisions for the existence of extraterrestrial intelligence; and that the confirmed existence of ET's will undermine seriously the already eroded credibility of Christianity as a way of knowing about the nature of reality.
After all, intelligent aliens are not animals, nor are they men. They certainly are not descended from Adam and Eve. Genesis, alone, precludes the very existence of such persons.
In the past, the church has modified its doctrine in order to accomodate contrary, irrefutable new knowledge about the nature of reality, such as the actual shape of the earth (though not, in this particular case, before murdering a number of heroically intransigent scientists). Some would no doubt aver that this points to a certain 'flexibility' in churh thinking that should not escape admiration; in my own view, however, it demonstrates that church doctrine is so subject to interpretation as to have no meaning, whatever.
That's my point. Of course, YMMV: that's why they print menus :)
Let's hope so!
Plus, none of our new friends will even have heard of Microsoft! I can't wait!
Yeah, and everyone should go to Australia and release non-indigenous toads. Not! How about restoring ecosystems, instead?
Smart servlet engines can detect when a user is refusing cookies, and do URL rewriting, instead. (don't know whether asp, cold fusion, &c do this).
How about a browser feature that refuses all cookies, and autosends a message to the originating site admin telling her to switch to jakarta-tomcat ;)
A person who can visualize regions of electron density the way most people can visuzlize legos has achieved mastery. A person who can merely rattle off constants has not.
In order to attain mastery of a subject, one must struggle through lots and lots of details, and undergraduate exams are mostly about testing a student's grasp of these details. But the ultimate goal of advanced education is to produce masters.
I would even go so far as to say that most undergrads needn't worry overly about committing all of these details to memory because they do not yet know which of them they will actually use in their careers (graduate or otherwise). When they start doing real work, they will naturally commit certain oft-used facts to memory. Others, they will just as naturally forget, and never miss them.
What matters in undergraduate education (other than beer ;) is that students gain sufficient background knowledge to be able to understand complex problems, and begin to think about them in relevant ways. From this all things proceed.