Semantic Entailment

1
Semantic Entailment

• Nathaniel Story
• Ginger Buckbee
• Greg Lorge
• Billy Dean

2
What is it?

• Given sentence A, can you infer sentence B?

iTunes software has seen strong sales in Europe. Strong sales for iTunes in Europe True
Kerry hit Cheney hard on his conduct on the war in Iraq. Kerry shot Cheney. False
3
Challenges

• Paraphrasing
• Negation
• Pre-Suppositions
• World Knowledge
• Juiciness

4
Paraphrasing Example

• There is a cat on the table.
• A cat is on the table.
• Different structurally, but infers same meaning

5
Negation

• I am lazy
• I am not lazy
• Im not unhappy (Double negation)
• Im happy
• Its not unnecessary
• Its necessary

6
Pre-Suppositions

• Bob doesnt think its raining
• Bob doesnt know its raining
• Conversational Pragmatics
• Contextual knowledge

7
World Knowledge

• Japan is the only country that currently has an
  emperor.
• Columbia doesnt have an emperor.
• First sentence entails second, but you need to
  know that Columbia is a country.

8
Approach

• Tools
• Stemmer
• Parser from Dan Bikels site
• MALLET (maxEnt classifier)
• Wordnet (synset)
• Focusing on Comparable Document task
• Start with simple features like word matching,
  synonym matching
• Add in more complicated functions like phrase
  structure comparisons
• Test the system out, see how it works. Continue
  adding features to improve performance.

9
Data

• Recognizing Textual Entailment Challenge (RTE)
  training data set
• Training set is labeled
• Best data set as was used in the European
  Competition

10
Evaluation

• International Competition
• Best 60 accuracy
• Strive for gt52 accuracy
• Comparing against annotated test set
• Improvement Print out incorrect ones, then look
  for mistakes.

11
The End
Questions?