Feeds

Potty-mouthed Watson supercomputer needed filth filter

Urban Dictionary blew its digital mind

Choosing a cloud hosting partner with confidence

IBM's Watson supercomputer was smart enough to beat two human opponents on US quiz show Jeopardy!, but there is apparently some knowledge that the system is still too immature to handle – namely, the contents of the Urban Dictionary.

Watson is perhaps the most sophisticated artificial intelligence computer system developed to date. In addition to its quiz-show triumph, it has already been put to work on a variety of real-world business scenarios, including analyzing customer data for the finance industry and helping doctors diagnose and treat patients.

Part of what makes Watson's problem-solving ability seem so uncanny is its advanced natural language processing capabilities, which allow it to understand the complexities of unstructured data. But as Fortune reports, IBM researcher Eric Brown learned that there are limits to how well Watson can interpret human communication.

In an effort to broaden the range of texts that Watson could interpret, Brown tried teaching it the Urban Dictionary, that crowdsourced compendium of words that tend to totter on the bleeding edge of modern English speech. WCGW?

Brown soon found that while Watson is fast with facts, it's none too quick to pick up on the subtleties of slang. Worse, like any schoolchild, the 6-year-old Watson was all too eager to try out the new words it had learned.

Almost immediately, Watson began casually dropping profanity into its everyday speech, such as answering one researcher's query with the less-than-scientific term "bullshit." (Although, to be fair, Watson may be sharper than the IBMers care to admit – just not particularly tactful.)

In the end, Brown and his team had to develop a filter to restrain Watson's more Tourette's-like impulses while they worked to undo the corruption that the introduction of the Urban Dictionary had caused.

Fortunately, Watson's current applications on Wall Street and in hospitals don't generally require it to know the difference between a ROFL and a WCETHBAC. But the incident does help to illustrate the fact that, as advanced as Watson may be, science is still a long way from developing a computer that can communicate as fully and richly as humans can.

"As humans, we don't realize just how ambiguous our communication is," Brown says, noting that Watson picked up a few bad habits from reading Wikipedia, too. ®

Internet Security Threat Report 2014

More from The Register

next story
Just don't blame Bono! Apple iTunes music sales PLUMMET
Cupertino revenue hit by cheapo downloads, says report
The DRUGSTORES DON'T WORK, CVS makes IT WORSE ... for Apple Pay
Goog Wallet apparently also spurned in NFC lockdown
IBM, backing away from hardware? NEVER!
Don't be so sure, so-surers
Hey - who wants 4.8 TERABYTES almost AS FAST AS MEMORY?
China's Memblaze says they've got it in PCIe. Yow
Microsoft brings the CLOUD that GOES ON FOREVER
Sky's the limit with unrestricted space in the cloud
This time it's SO REAL: Overcoming the open-source orgasm myth with TODO
If the web giants need it to work, hey, maybe it'll work
'ANYTHING BUT STABLE' Netflix suffers BIG Europe-wide outage
Friday night LIVE? Nope. The only thing streaming are tears down my face
Google roolz! Nest buys Revolv, KILLS new sales of home hub
Take my temperature, I'm feeling a little bit dizzy
prev story

Whitepapers

Choosing cloud Backup services
Demystify how you can address your data protection needs in your small- to medium-sized business and select the best online backup service to meet your needs.
Forging a new future with identity relationship management
Learn about ForgeRock's next generation IRM platform and how it is designed to empower CEOS's and enterprises to engage with consumers.
High Performance for All
While HPC is not new, it has traditionally been seen as a specialist area – is it now geared up to meet more mainstream requirements?
Saudi Petroleum chooses Tegile storage solution
A storage solution that addresses company growth and performance for business-critical applications of caseware archive and search along with other key operational systems.
Simplify SSL certificate management across the enterprise
Simple steps to take control of SSL across the enterprise, and recommendations for a management platform for full visibility and single-point of control for these Certificates.