New tasks emerging

Now we will investigate several novel areas that have emerged in the recent past. These areas include detecting sarcasm, language grounding (that is, the process of eliciting common sense from natural language), and skimming text.

Detecting sarcasm

Sarcasm is when a person utters something which actually means the opposite of the utterance (for example, I love Mondays!). Detecting sarcasm can even be difficult for humans sometimes, and detecting sarcasm through NLP is an even harder task. Sarcasm SIGN: Interpreting Sarcasm with Sentiment Based Monolingual Machine Translation [23], Lotem Peled and Roi Reichart, uses NLP for detecting sarcasm in Twitter posts. They first create a dataset of 3,000 tweet pairs, where one tweet is the sarcastic tweet and the other tweet is the decrypted nonsarcastic tweet. The decrypted tweets were created by five human judges who looked at the tweet and came up with the actual meaning. Then they used a monolingual machine translation mechanism...