Later on in this course, we will show how
natural language processing techniques can
be used to extract deeper information from
human language.
Let's now recap what we've learned this
week.
We actually started with the information
theory from Shannon that was invented for
a totally different context, looked at the
concept of mutual information and applied
it to the statistics of language, showed
how we could summarize documents and
discern the best possible keywords using
TF-IDF which turned out to be exactly the
same thing as mutual information.
We figured out the relationship between
communication and machine learning in
terms of mutual information learning
again, and learned the very important
nieve base classifier, which is the
foundation for almost all machine learning
techniques.
Then we summarized with the limits of
machine learning, both from the
information theoretic perspective which
also told us which features to use and
which not to use.
And lastly we ended with some suspicions
about whether the bag of words approach
that we'd used just considering without
their grammatical syntax or semantics was
actually enough to discern meaning.
In future class we will ask questions such
as where features themselves come from.
For the moment, we have chosen features
like words and we have labeled passed data
manually or by experience, such as buyers
and browsers.
In our lives however, the labels and the
features need to be derived automatically
by us as we learn about the world with no
supervision or nobody telling us what's
the feature and what's not.
Before we come to those very interesting
ideas in the world of learning, we'll
first take an excursion into big data
technology next week as we have promised.
We'll describe how the new technologies
that were developed in the web world
differs significantly from traditional
technologies.
And then we'll do some experiments and
assignments of how they can be used for
indexing, page rank, computing TF-IDF,
implementing naive Bayes classifiers,
computing mutual information, and all the
nice stuff that we have learned so far,
including locality sensitive hashing, that
we did last week.
We've learned a lot of theory, done some
calculations.
And now, get ready for doing some
implementation, And programming.
So see you next week.
And of course, don't forget to submit your
homework by Monday.