Jump to content
Washington Football Team Logo
Extremeskins

I Smell a Nobel Prize - "That's What She Said" Software Developed


Dan T.

Recommended Posts

Those who thought that the departure of Michael Scott from The Office would result in a shortage of "That's What She Said" jokes, fear not. Two University of Washington graduate students have developed a computer program that can analyze language in order to crack That's What She Said jokes.

The program, called the Double Entendre via Noun Transfer, or DEviaNT, developed by Chloé Kiddon and Yuriy Brun, recognizes nouns that can have double meaning in sentence structures that are ripe for TWSS punch lines.

Their paper cites several successful DEviaNT examples, including responses to "Yes give me all the cream and he's gone," and "Don't you think these buns are a little too big for this meat?"

Their algorithm was constructed to minimize false postives (though a bad TWSS joke might still be funny if told by a computer), With tinkering their program has achieved a 71.4 percent success rate.

Here's their research paper: http://www.cs.washington.edu/homes/brun/pubs/pubs/Kiddon11.pdf

Enjoy, even though it's a little dry.

Link to comment
Share on other sites

What a great use of research dollars... wonder if they are on grant and who funded them.

Looks like I found who:

This material is based upon work supported

by the National Science Foundation Graduate Research Fellowship under Grant #DGE-0718124 and

under Grant #0937060 to the Computing Research

Association for the CIFellows Project.

Ugh ... more ammo for the Republicans who want to cut NSF.

Link to comment
Share on other sites

What a great use of research dollars... wonder if they are on grant and who funded them.

Looks like I found who:

Ugh ... more ammo for the Republicans who want to cut NSF.

Disagree. It may seem trivial, but their work could help refine computers' ability to decipher spoken language. It's bigger than it looks at first glance.

Link to comment
Share on other sites

Disagree. It may seem trivial, but their work could help refine computers' ability to decipher spoken language. It's bigger than it looks at first glance.

They've been doing stuff like this for 30+ years ... look up ELIZA ... computers have no inherent concept of what most words mean since they don't have our sensory experience ... as compelx as these programs can get, they are like parlor tricks.

Link to comment
Share on other sites

They've been doing stuff like this for 30+ years ... look up ELIZA ... computers have no inherent concept of what most words mean since they don't have our sensory experience ... as compelx as these programs can get, they are like parlor tricks.

Parlor tricks with potentially useful business or medical applications. Those tricks should pay off.

Link to comment
Share on other sites

:)

-------

That's what she said was used so frequently in my apartment Junior/Senior year of college that we just resorted to laughing when it was appropriate to say it.

That sounds like a rollicking good time.

Link to comment
Share on other sites

They've been doing stuff like this for 30+ years ... look up ELIZA ... computers have no inherent concept of what most words mean since they don't have our sensory experience ... as compelx as these programs can get, they are like parlor tricks.

Computers have no inherent concept of what most words "mean" because they have no subjective consciousness, hence there is no way for them to self reference words with the experience and ultimately the exchange of that experience (using symbols, words) with another being also posessing a subjective consciousness and self awareness.

But fear not. You will bow before the new AI overlords* in the near future. ;)

*...or hypnotoad

Link to comment
Share on other sites

people don't have an inherent concept of what words mean, either. People learn through immersion, and computers (usually) learn through programming.

A computer can "learn." You just have to drill it over and over again.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...