Question 1. After signing up on https://piazza.com/umd/spring2018/ling773cmsc773inst728c/, please complete this Google Survey. Although this assigment is not due until next class, it would be really helpful to get responses to the survey as soon as possible, ideally before the end of the weekend, since they could influence planning for upcoming lectures.
Here's an example of an answer if we were asking about coreference:
Coreference involves the use of terms that refer to other terms in an utterance or conversation, typically entities that have already been referrred to. Often this is accomplished using pronouns (e.g. "The professor gave out his first
assignment and it was really interesting), but definite references (e.g. "that professor") can also be used to co-refer. Here are some things we could say to Alexa to explore her co-reference capabilities:
(Alexa answers)
Alexa, when did he die?
This will test whether Alexa handles (some) coreference to named entities. We could also ask "When did she die", to see whether coreference is paying attention to pronoun gender distinctions (masculine vs. feminine), "When is his birthday", to see whether she handles reference using a possessive pronoun, or "When did that president die" to see whether she handles definite reference.
Alexa, when was George Washington born?
Please don't forget to tell us approximately how much time you spent on this question.
Question 3. Recall that Bayes's Rule is often used in "Bayesian updating", since it is used to take a prior probability distribution and update it on the basis of observed evidence to produce a posterior distribution. Before, after, hence updating; get it?
This notion of "Bayesian updating" is particularly useful when applied iteratively. Consider a scenario where we have a prior distribution P(H) over hypotheses, and then we observe evidence E1 and use Bayes's Rule to update our distribution to
P(E1 | H) P(H)
P(H | E1) = --------------
P(E1)
The distribution P(H|E1) can now be viewed as a new, more informed
prior. Therefore if we see a new piece of evidence E2, we can use
Bayes's Rule again to derive a new posterior. Since P(H|E1) is the
prior now, the result will take
the form
X P(H | E1)
P(H | E1,E2) = ------------
Y
Derive the simplest possible form of the full expression, i.e. fill in X and Y, assuming that E1 and E2 are independent. (Note that if you don't make that assumption, you won't get the right answer.)
Please don't forget to tell us approximately how much time you spent on this question.
Then, go to Wolfram Alpha and try typing the following things into the box at the top (and then hit Enter or click the "=" sign).
4(a) Look at the PDFs (probability density functions) for each one, and explain in plain English what each of these Beta distributions represents, when interpreted as a Bayesian prior for the probability of heads. (To keep things standard, let's all assume that the prior belief is about the probability of heads for a possibly-unfair coin.) What is your expectation about what the probability of heads will be, and how strongly do you think so?
4(b) Let's take the notion of "expectation" in (4a) a little more literally, as in expected value, also known as the mean of the distribution. Try typing BetaDistribution[α,β] for some other values of α and β, and look at the number Wolfram Alpha reports as the "mean". Can you figure out what the formula for the mean is, as a function of α and β? (If you can't give an exact formula, at least comment on some properties of the mean. If you already knew the formula, it's ok to tell me so. But don't just look up the mean of the Beta distribution and copy it down; what would you learn from that?!)
Please don't forget to tell us approximately how much time you spent on this question.
Please do not include your code in the same document as your explanation. If we print out your writeup we don't want to be printing out code! Include it as a separate attachment when you turn in your assignment.
Note that the goal here is not to impress me with how well you can write a program to extract bigrams; see the definition of quick and dirty if you weren't already familiar with that expression. Using existing code is ok; if you can find software somewhere with a function called
extract-nonstoplist-bigrams-from-a-web-page-and-sort-by-frequencyby all means go ahead and use it. Ok, that probably doesn't exist, but if you want to use an existing toolkit, that's fine. (If you code in python, two good existing packages are NLTK or spaCy. If you are choosing between those two, NLTK has the advantage of being very tutorial, including having an associated textbook, but spaCy is is what you'd want if you were writing "production" software.) Like the other questions on this assignment, this is a reality check: if you find it very challenging or time consuming, then we should probably talk about whether this course is right for you.
Please don't forget to tell us approximately how much time you spent on this question.
Please don't forget to tell us approximately how much time you spent on this question.
Please don't forget to tell us approximately how much time you spent on this question -- in this case how much time you spent on the reading. Note that in general, a good strategy for this class is to do a relatively quick read-through of the assigned material, take notes in lecture, and then go back and re-read with better knowledge of what I was focusing on and what you do and don't already understand.