Using conceptual dependency: Write CD representations for the following sentences. Unfilled slots in the CD representations can either be left in (to make the missing information more obvious) or omitted for convenience (CD-based programs would always know they're unfilled, though). Because we will not address object representations, try to make up your own scheme for representing objects, analogous to CD's representation for actions. If you notice an ambiguity that must be resolved before the representation can be done, point out the ambiguity and how you are resolving it.

1.�� Sally drove a motorcycle to New York.

(ptrans(actor person (name (Sally)))

(object person (name (Sally)))

(to New York)

(instrument motorcycle))

 

2.�� John donated his sofa to the Salvation Army.

(atrans(actor person (name (John)))

(object sofa)

(to Salvation Army))

3.�� John roared into town in a new BMW. (For this sentence, represent reasonable inferences.)

(ptrans (actor person (name (John)))

(object person (name (John)))

(to town)

(instrument BMW))

4.�� John ate popcorn with his foot.

(ingest(actor person (name (John)))

(object popcorn)

(instrument foot))

5.�� Mary broke John's TV.

����������� (propel (actor person (name (Mary)))

����������������������� (object tv (John)))

 

 

Ambiguity of words: As the previous example shows, MicroELI can handle ambiguous verbs fairly easily. Discuss the issues involved in writing requests that would enable mcELI to handle ambiguous nouns. (E.g., the word ``bill'' in ``Bill paid the bill.'') Sketch how you might change the MicroELI framework and how you would write requests using the new framework to enable the program to parse the sentence ``Bill paid the bill with a bill.''

 

Ambiguous nouns are given meaning through the context in which they are used.mcELI is able to parse a sentence like �Bill paid the bill� by looking at �paid� as an atrans of money from the actor, or subject of the sentence.Thus, the first occurrence of �Bill� would be given as a name.In the sentence �Bill paid the bill with a bill� would be handled in the same way, except when the atrans for �paid� filled in the information, it would look for a noun for the subject and a noun-phrase for the method.While this would work in the two examples here, there would still be problems in using the new framework.

 

 

Difficulty of CD: Spend some time thinking about what makes a sentence hard or easy to represent using CD. What kinds of sentences are easy to represent? What kinds of sentences are hard? Give concrete examples of sentences illustrating the types of issues you identify. Why?

 

Sentences that are easy to represent using CD are sentences that have only one meaning or idea in them and represent an action.This is because CD was created to represent simply sentences, even if they hold two separate ideas that are contradictory (like the last example: treat it like two sentences).

����������� Sally ate the bread and butter.

����������� John thought about his girlfriend.

����������� Sammy wanted to drink a Coca-Cola, but he bought a Pepsi instead.

 

Sentences that are difficult to represent using CD are sentences in which a fact is stated.Instead, the fact should be represented differently, so that the information can be used later.

����������� John has a girlfriend.

����������� Sally is John�s girlfriend.

 

Other sentences that are difficult to represent using CD include sentences where more information is stored using a complex sentence structure.While parts of the sentence can be represented, it does not properly convey the meaning of the sentence.Also, appositives make it difficult to decide what to put into the CD representation

����������� While John was thinking about his girlfriend Sally, Sally was making out with Sammy in the backyard of John�s house.

 

Sentences that are difficult to represent using CD can also be negatives because there is no standardized way of showing it, unless it is written into the CD by the user.

����������� Sally is not faithful to John.

����������� John does not know Sally is playing the field in his backyard.

 

 

A Problem: Assume that the sentence ``A coin rolled into the store'' is analyzed as (PTRANS (OBJECT (MONEY)) (TO (STORE)))  Process this line with micro-SAM. Then process the CD representation for ``John bought a restaurant.''

1.      Explain what goes wrong in each case and the causes of the problems.

In each case, the problem is that the key words in the sentence (store and restaurant) call up the stories for going to a store or restaurant.Thus, in the first case, there is no actor, simply the store.Thus, there is a story with no one in it.In the second case, John purchases a restaurant, not a meal in the restaurant.Thus, the restaurant becomes a meal with in the story.

2.      What is a simple way that you could modify the code to address this? (You do not need to write this code; simply give your strategy.)

A simple way to modify the code would be to look for the first sentence structure of the story to be the first sentence given as an argument.If the first sentence structure was found, it would give the entire story; otherwise it would return the given sentence.

3.      When might your simple approach fail?

This approach will fail when the beginning sentence is not the first sentence in the script.One case is the sentence �John ordered steak.John paid the waiter money.He left the restaurant.�The approach above would not recognize this as John going out for dinner.Instead, it would be regurgitated.

4.      What would you propose to address this problem? Try to think of a method that is robust (applies to a wide range of situations) and efficient, and sketch its tradeoffs. (Addressing this problem in a rich way has been the subject of much work in the literature, as well as the heart of at least one dissertation in AI.)

The best way to address the above problem is to compare the sentences given with the sentence structure of the stories.If the sentence structure matches, then it most likely fit into the story.Of course, this can take a long period of time for all the comparisons.A way of implementing this is to complete all the stories, and then look at the blanks within the stories.The story with the fewest blanks would be the correct one, with there being a threshold requirement, so if none fill the blanks properly, then the sentences are taken without a script.