Frequently Asked Questions (WordNet)

Can I read the synset or hypernym file twice? No. File I/O is very expensive; read each file only once and store it in an appropriate data structure.

Any advice on how to read and parse the synset and hypernym data files? Use the readLine() method in our In library to read in the data one line at a time. Use the split() method in Java's String library to divide a line into fields. You can find an example using split() in Domain.java. Use either Integer.parseInt() to convert String id numbers into int values or Integer.valueOf() to convert into Integer objects.

Which data structure(s) should I use to store the synsets, synset ids, and hypernyms? This part of the assignment is up to you. You must carefully select data structures to achieve the specified performance requirements.

What should sca() return if is there is a tie for the shortest commmon ancestor? The API does not specify, so you are free to return any shortest common ancestor.

Do I need to store the glosses? No, you won't use them on this assignment.

What is the root synset for the WordNet DAG?

38003,entity,that which is perceived or known or inferred to have its own distinct existence (living or nonliving) 

Can a noun appear in more than one synset? Absolutely. It will appear once for each of the noun’s distinct meanings. For example, the noun word appears in these 8 synsets:

35532,discussion give-and-take word,an exchange of views on some topic; "we had a good discussion"; "we had a word or two about it"
56587,news intelligence tidings word,new information about specific and timely events; "they awaited news of the outcome"
59267,parole word word_of_honor,a promise; "he gave his word"
59465,password watchword word parole countersign,a secret word or phrase known only to a restricted group; "he forgot the password"
81575,word,a string of bits stored in computer memory; "large computers use words up to 64 bits long"
81576,word,a verbal command for action; "when I give the word, charge!"
81577,word,a brief statement; "he didn't say a word about it"
81578,word,a unit of language that native speakers can identify; "words are the blocks from which sentences are made"; "he hardly said ten words all morning"
Do not assume that the number of synsets in which a noun participates is bounded by a constant. The noun head appears in 33 synsets.

Can a synset consist of exactly one noun? Yes. Moreover, there can be several different synsets that consist of the same noun.

62,Aberdeen,a town in western Washington  
63,Aberdeen,a town in northeastern South Dakota  
64,Aberdeen,a town in northeastern Maryland  
65,Aberdeen,a city in northeastern Scotland on the North Sea  

I'm an ontologist and I noticed that your hypernyms.txt file contains both is-a and is-instance-of relationships. Yes, you caught us. This ensures that every noun (except entity) has a hypernym. Here is an article on the subtle distinction.

Frequently Asked Questions (ShortestCommonAncestor)

Can I use my own Digraph class? No. You must use Digraph.java. That is the type of the argument to the constructor.

How can I make the data type ShortestCommonAncestor immutable? You can (and should) save the associated digraph in an instance variable. However, because Digraph is mutable, you must first make a defensive copy by calling the copy constructor in Digraph.

Is a vertex considered an ancestor of itself? Yes.

What should ancestor() return if is there is a tie for the shortest common ancestor? The API does not specify, so you are free to return any shortest common ancestor.

I understand how to compute the length() method in time proportional to E + V in the worst case but my lengthSubset() method takes time proportional to a × b × (E + V), where a and b are the sizes of the two iterables. How can I improve it to be proportional to E + V? The key is use a multi-source version of breadth-first search, as in the the constructor in BreadthFirstDirectedPaths that accepts an iterable of sources as an argument (instead of a single source).

Should I construct a new ShortestCommonAncestor object for each call to sca() and distance()? No. You need only one ShortestCommonAncestor object per WordNet object. The methods sca() and distance() should make one call to either lengthSubset() or ancestorSubset().

For the extra credit, do length(), lengthSubset(), ancestor(), and ancestorSubset() need to take time proportional to to the number of vertices and edges reachable from the argument vertices in the worst case? Or, may I use hashing? You can make standard technical assumptions (such as the uniform hashing assumption). If you do so, state any assumptions that you make in your readme.txt file.

Should I re-implement breadth-first search to compute shortest common ancestors? Not unless you are attempting the extra credit. For most students, we recommend using BreadthFirstDirectedPaths.java.

Should I re-implement depth-first search to find directed cycles? No. We recommend using either DirectedCycle.java or Topological.java.

Do I need to throw exceptions explicitly with a throw statement? No, it's fine if they are thrown implicitly, e.g., you can rely on any method in Digraph.java to throw a java.lang.IllegalArgumentException if passed a vertex argument outside of the prescribed range. A good API documents the requisite behavior for all possible arguments but, hopefully, you should not need much extra code to deal with these corner cases.

Frequently Asked Questions (Outcast)

What should outcast() return if is there is a tie for the outcast? The API does not specify, so you are free to return any outcast.

My algorithm computes the distance between every pair of nouns. Is that okay? Yes, that's fine.

Input, Output, and Testing

Some examples. Here are some interesting examples that you can use to test your code.

Possible Progress Steps

A video is provided for those wishing additional assistance. Warning: the video was made in early 2014 and is somewhat out of date. For example the API has changed.

Enrichment