NLP Notes

T-H pair'leri asagidaki yarismalarin dataset'lerinden turetilmis.

Information Retrieval:
- Select a news title. (H)
"The U.S. military evacuated U.S. citizens."
- Search it in a search engine.
- Find the best item. (T)
"A total of 599 United States citizens were evacuated safely, at their request; those who were interviewed expressed great relief at being out of Grenada."

Comparable Documents:
- Obtain a cluster of comparable news articles.
- Examine "aligned" sentence pairs that overlap lexically, in which semantic entailment may or may not hold.
- Pick them as T-H pairs.

Reading Comprehension:
- Read a news article.
- Make a particular assertion from the text.
- Create a T-H pair.

Question Answering:
- Pick a question.
- Answer the question yourself (H)
- Ask it to a QA system.
- Select the best text snippet that includes the answer (T)

Information Extraction:
- Pick a relation of interest. "killings of civilians"
- Search it in a IE system.
- Find the best hit (T)
- Turn the relation of interest into a sentence (H)

Machine Translation:
- Pick a text.
- Get the human translation.
- Get the machine translation.
- Pick T-H pairs.

Paraphrase Acquisition:
- Pick a sentence. (T)
- Get a list of paraphrases from a PA system.
- Pick the best paraphra (H)

For more information:

The PASCAL Recognising Textual Entailment Challenge: http://www.cs.biu.ac.il/~glikmao/rte05/dagan_et_al.pdf

RTE1 Dataset / Development: http://www.cs.biu.ac.il/~glikmao/dev2.zip
RTE1 Dataset / Test: http://www.cs.biu.ac.il/~glikmao/test.zip
RTE1 Dataset / Annotated: http://www.cs.biu.ac.il/~glikmao/annotated_test.zip

RTE 2 Dataset: http://ir-srv.cs.biu.ac.il:64080/RTE2/PASCAL/index.php?action=download