Teaching language to a machine

We posted last week about the Children of the Code Web site, and noted that reading is such an incredibly complex task that it’s not notable that some students struggle with reading, but rather miraculous that any of us can read at all. Computers are good at breaking down complex tasks like forecasting weather – could they be any good at learning language and reading?

Today’s NY Times describes a computer program under development at Carnegie-Mellon University called NELL (for Never Ending Language Learning). NELL is attempting to learn by acting not like a computer (computers are generally very good at following rules – for example, learning to play chess – but lousy at more nuanced tasks), but like a human being.

Researchers working on NELL cited an example of the following two sentences:

The girl caught the butterfly with the spots.

The girl caught the butterfly with the net.

A human reader inherently understands that girls hold nets, and girls are not usually spotted. So, in the first sentence, “spots” is associated with “butterfly,” and in the second, “net” with “girl.”

“That’s obvious to a person, but it’s not obvious to a computer,” Dr. Mitchell said. “So much of human language is background knowledge, knowledge accumulated over time. That’s where NELL is headed, and the challenge is how to get that knowledge.”

But if a computer is using a hierarchy of rules self-developed rules to resolve ambiguity in language, what happens if it gets a rule wrong?

When Dr. Mitchell scanned the “baked goods” category recently, he noticed a clear pattern. NELL was at first quite accurate, easily identifying all kinds of pies, breads, cakes and cookies as baked goods. But things went awry after NELL’s noun-phrase classifier decided “Internet cookies” was a baked good. (Its database related to baked goods or the Internet apparently lacked the knowledge to correct the mistake.)

NELL had read the sentence “I deleted my Internet cookies.” So when it read “I deleted my files,” it decided “files” was probably a baked good, too. “It started this whole avalanche of mistakes,” Dr. Mitchell said. He corrected the Internet cookies error and restarted NELL’s bakery education.

The researchers behind NELL (and other projects that are attempting to teach computers to attack language as humans do) cite the possibilities for improved natural language search (where searching returns answers to questions, rather than just lists of relevant Web sites) as a positive outcome of their research. One hopes as well that as we train a computer to think like a human we gain additional insight into how humans think and learn, with the potential to improve learning for our children.

Tags: , , ,

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: