h1

learning 2.0

October 4, 2010

If you take a quick glance, this New York Times science article (available here) may seem to have little to do with psychology – a computer assessing information on the web to teach itself about the world? Why would someone interested in human behavior have time for such a topic? I guess there are a couple of reasons.

First, psychologists have always relied on convenient metaphors to convey there ideas (e.g. the “iceberg metaphor” to describe Freud’s theory of the unconscious or the ” hydraulic system metaphor” to explain leakage from the unconscious affecting behaviors). In the last quarter of the last century, one of the most prevalent metaphors in psychology, thanks to the cognitive renaissance, was the computer metaphor of the mind. We have a brain (hardware) and various cognitive processes (software) that enable thought and thus behavior. There has been a shift away from this metaphor in the last decade because it is too simplistic, and the human mind doesn’t operate like a computer in many, many ways. However, the development of dynamic, computational learning systems may create new metaphors that can be used to think about human cognition.

Second, cognitive psychologists have a long history of working with computer scientists in an effort to (a) understand the various factors that influence learning and (b) create computational models that capture important capture important qualities of cognitive representations, processes, and systems (e.g. John Hummel’s work). As we see how NELL learns, or fails to at particular points, we may gain insight into why people have particular difficulty, or ease, with some learning situations. I thought the passage from the article below nicely captures one of those kinds of moments – by understanding NELL’s hang-up we can reflect on our own:

“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.”

This is the same confusion I sometimes see in students when we get part way into a new topic – everyone is fine and seems to be learning the new material… but then it all falls apart. Usually it has happened that one simple misunderstanding cascades into a complex and confusing mess. Not so different from NELL’s hiccup with internet baked goods…

So, I guess in the end, we should pay attention to efforts to build these learning machines, if only so that we can take solace in the fact that even though learning is a horribly complex and difficult task on some level, we are able to succeed in acquiring new knowledge with relative ease. But that is no excuse for not studying for the upcoming exam…

Advertisements

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: