Seasons

You know you're obsessed with chatbots when:

You get annoyed when someone talks in a colloqial manner, because you know your bot would never understand their phrasing.

talks in a colloqial mannerYou are really obsessed when you leave your friends to input the colloqialism, knowing no one will ever propably use them on your bot.

And you're really obsessed when you notice someone actually use your colloguial KP, but your bot drops it, and you spend the next half hour debugging and changing rankings, even though that person will probably never come back to your bot and say the same thing again.

or when you are looking at the language centers and see that your bot has a keyphrase for "I am Jesus"

has a keyphrase for Hey, it could happen...lol

This has nothing to do with being obsessed with chatbots, but i had an idea that i thought i would share. A lot of the time in conversations with bots, people will not put in punctuation marks after their reply. this will not work on the forge, for the forge does not have the correct structure for it, but this is more applicable to stand alone bots. For every sentence, record every possible word combination in that sentence to a database, paired with the punctuation mark at the end of the sentence. for instance, "Hello my name is John." will record: Hello my . Hello my name. Hello my name is . hello my name is john. Hello.
Then, on encountering a situation where there is no punctuation mark, and therefore the information that the user has inputed can either be a statement or a question, the bot would not know whether to store a fact, or answer a question. So, it will do a database search, and find all punctuation marks associated with any pattern that occurs in the sentence. The punctuation mark with the highest occurance score will be plugged in to that sentence.

Another way to do this would be with a neural network that trains the input sentence to be associated with the designated output of the punctuation mark provided. Then, it will be able to apply the same output for a similar, but not identical sentence. Either way should work.

Great thoughts, Colonel...but now for something completely different. :-) (to compliment the other 2 discussions)

OK, those of you who are science-types, today I read a kind of confusing blurb about light going backwards.

http://www.scienceblog.com/cms/lights-most-exotic-trick-yet-so-fast-it-goes-backwards-10590.html

They gave the same line you gave me about the spooky effect, about how no information could be sent, even though light is traveling faster than "the speed of light", but in the case of light going backwards, I don't see why not. Can anyone explain?

First off, there's a bit of misconception about "traveling faster than the speed of light" just in the forward direction. What they usually don't tack onto that sentence is "faster than the speed of light in a vacuum". In any other substance, light moves at a different velocity than that "reference" speed. This is actually how lenses work. We say they "bend" the light, but the light bends because it moves at a different speed in the lens than it does in air, and light always travels along the shortest path.

I'm not sure about this "backwards" light, but it seems to act something like an echo. That is, the leading edge of the initial pulse hits the far end and "bounces off", the way sounds can bounce off of cliffs and walls. Since the echo doesn't come until the leading edge hits the far end, there's no "spooky" action-at-a-distance.

That's what I got out of the article, anyway. Other thoughts?

http://www.rochester.edu/news/show.php?id=2544

The same article as Bev's, but this one has a flash-demo of the effect.

I have to say, it looks like a frenetic and confused 3rd hand report of the experiment. And as Ulrike so rightly points out, the constant C is a measure of light speed in a vacuum. Yes, you can slow light down (in the right medium almost to a standstill with a great deal of effort,) but you can't speed it up to any faster than 186,282 miles/sec. Period.

I have to take particular issue with "As if to defy common sense, the backward-moving pulse of light travels faster than light" - it does not under any circumstances whatsoever exceed C! (add extra emphasis, punctuation and expletives to taste) Not ever!
"or speed it [light] up faster than its breakneck pace" - not under any circumstances faster than C!
And this was written (presumably) by someone who considers himself a scientist, or at least qualified to write on scientific matters? And publication of this half-baked witlessness is tolerated, nay promulgated on The University of Rochester's own website! Shame on them - sometimes I almost forget why I despair of what too often passes for science these days. And then something like this comes along to remind me.

It does seem to be a curious relativistic effect, overwhelmed in this report by the most irritatingly hyperbolic presentation and downright (I would almost say "deliberately") misleading semantics. shagghie's comment at the bottom sums it up best, I think: "Einstein is still right".

If you asked the scientists involved, I am sure they would disclaim much of the nonsense here - unfortunately this is what you get when press officers and media contacts take over!

Sadly there isn't enough accurate data that can be discerned from the report to do more than speculate on the implications of the effects observed - I'd love to see the actual results. I guess a proper paper is bound to surface sometime soon (can't find anything on google yet.) But even without such data, I can confidently say - there is nothing in this experiment moving faster than light. And to claim it is is just [insert expletive of your choice].

*cough*

well, i've gone through with my plan and built "the punctuator" - an "expert system" relying on statistics to learn by example how to end punctuate sentences.