[MUSIC] There are many systems for
restaurant reviews out there but today we're gonna talk about a pretty
new and exciting one that we can build. So it's an important day for us and we're trying to book
a Japanese restaurant, and we love Japanese food and want a really
cool place to have some Japanese food. But this is Seattle that we live in,
and Seattle is really awesome for Japanese food, so there are many,
many places for sushi, and in fact, there are many that have really
good ratings, say four stars. So perhaps when I'm thinking
about a restaurant to go to I'm not just thinking about the overall
stars of the restaurant, but I'm thinking about how is it in terms
of the food, in terms of the ambiance? In particular how is it
in terms of the sushi, which is the thing that
I really care about. I really love sushi and I want really,
really the best, freshest fish, and the most awesome,
innovative sushi I can get. So when you talk about a positive review
for a restaurant, and here's a restaurant that we love in Seattle, a sample
review might go something like this. So watching the chefs
create an incredible, edible piece of art,
was really a unique experience. So the experience is really positive if we
think about that aspect of the restaurant. But a review might want to say,
well, my wife tried the ramen, and it was pretty forgettable. That means the ramen, not so good. Now, the review continues and says, all the sushi was delicious,
easily the best sushi in Seattle. So the sushi was really two thumbs up. Now, I don't care about the ramen. I'm not going to this place for ramen. I don't care that this person
really didn't like the ramen, or his wife really didn't like the ramen. What I really care about is good
experience and amazing sushi. So this review is just for me. So when I'm thinking
about restaurant reviews, I wanna understand the aspects of
the restaurant, their positive or negative, and really think about which
ones of those really affect my interests. So if I look at all the restaurant
reviews, I want to feed them to this new, really exciting, new type of restaurant
recommendation system that's gonna tell me the experience was good, four stars. The ramen was a so-so, but who cares? But the sushi, the thing that I really
care about, that was five stars, and not just that, it's gonna give me
some interesting feedback, so for example, this is easily
the best sushi in Seattle. Now that's where I wanna go. So how do we build such an intelligent
restaurant review system? Well we're going to start
with all the reviews and we're going to break
them up into sentences. So each review is composed
of multiple sentences and some sentences cover different
aspects of the restaurant. So for example, a sentence that says
easily the best sushi in Seattle, I wanna feed that sentence
through a sentiment classifier. So sentiment classifier is gonna
say is this sentence positive, or is it a negative sentence? Now we have a sentiment classifier,
which can take a sentence and say is it positive or negative? So how do we build this new kind of
cool restaurant review experience? So we're gonna take all the reviews and break them into sentences
that will be discussed but then we're gonna select the sentences
that talk about the aspect I care about. So which ones of these
are really about sushi? So not all of them, but a subset of those. Then I feed those sentences
through the sentiment classifier, each one of them, and
then I average the results. In this case all of
the sentences were positive and so I'm gonna say this is a five star
restaurant review in terms of the sushi. Not just that, I can look at
the sentences about sushi and look at which ones are most positive. So for example, this one must be
the positive, but easily the best restaurant in Seattle maybe is the most
positive, and I can display those. I can also display the most negative ones. In this case, there's nothing
negative to say about the sushi. [MUSIC]