[MUSIC] So, we now built a whole new kind
of restaurant experience, or restaurant review experience,
using classifiers. Let's dig in and really understand
a little bit more what a classifier is, and some other applications
of classification. So a classifier takes some input x, for
example a sentence from the review or other inputs as we'll see. It pushes it through what's called
a model to output some value y, that we're trying to predict. And here it's a class, for
example, positive or negative. So positive in the case of sentiment
analysis corresponds with thumbs up reviews, while negative
corresponds with thumbs down. But this is just one
example of classification. You can look at text, for
example, a web page, I want to figure out which webpages interest me,
and I can align them to categories. For example, is this about education,
a page about education, is it a page about finance,
is it a page about technology and so on? So, there's not just two categories. There can be three, four, or even thousands of categories
I'm predicting from. Now, another example of classification, which really has impacted all of
our lives, is in spam filtering. Some of you might remember, perhaps in the
early 2000s, what spam filters were like. The quality was not very good. Really, they were all hand-tuned
things where somebody said, oh it has certain words, it must be spam. But the spammers kept changing
the words a little bit, adding numbers instead of letters,
and beating the spam filters. And what really changed
the world of spam filtering. It changed it so much that I don't even
look at my spam folder, actually, sorry if your message went to my spam folder,
I just don't open it, is machine learning. It's classifiers. They took the input x of the email and they fed it through a classifier that
predicted whether it's spam or not. It did that really well. And it does it by not just looking
at the text of the email, but it looks at other characteristics. So for example, who sent it. If it somebody's whose a close friend,
or somebody you have lots of vacation, with it's less likely to be spam. The IP address is a person sending
from the usual computer and so on. Lots of information. So this is another really
interesting practical application. In computer vision,
we do a lot of classification. We'll take an image, and
figure out what is in that image. For example, here,
the input x are the pixels of the image. We feed it to a classifier, and we're going to predict
things like is this a dog? In fact it's a Labrador Retriever, Golden
Retriever, or a different kind of dog. This is actually my dog,
Labrador Retriever. And as we will see later on
in the deep learning module, there's really interesting new ways of
doing this with very high accuracy. Now you can also use classification
in medical diagnosis systems, and in fact this is what your doctor does. They take your temperature, maybe they
look at your x-ray, they look at some medical tests and they try to make
a prediction about what's ailing somebody. Maybe they say nah, nah,
you're just healthy, or you have a cold, you have the flu,
maybe even you have pneumonia. That's a variable y
that's being predicted, what is the disease that's going on. Now these days, there's really interesting
new things around personalized medicine, because the prediction doesn't just
depend on the standard measurements, but can be really personalized for me. Can depend on my particular DNA
sequencing, which is pretty exciting and also my lifestyle,
which maybe look something like or maybe more realistic,
something like this. And so, given all these measurements, can make an even better
prediction at what's ailing me. Now this idea of classification
in off-machine learning, has really gone much further,
even to be able to read your mind. So when I was at Carnegie-Mellon, Tom
Mitchell, one of my friends and colleagues at his office next door, and he come to me
and says, we've done this amazing thing. Which would take an image of your
brain using a technology called FMRI, which is brain scan, and predict
when you're reading a word of text, whether you're reading the word hammer or
the word house. So it's really reading your mind. But in fact,
he went on to do many interesting things. For example, if you're looking at
a picture of a hammer or a house, but you train the classified on them you
reading words hammers and house, you're still able to read your mind and
figure out what picture you're looking at. So this is yet the next frontier of classification
to understand how the brain works. [MUSIC]