ProfessorQ Answers Your Travel-Related Questions

	Please type your question in here:

Summary

This technology is on responding to a natural-language question. One approach is to match the question with existing questions in a database, for example, by comparing their essential grammatical components. Another approach is to analyze the question based on grammatic and sematic rules, with additional features, in general, including:

Features Include:

• If a word in the question has more than one meaning, provide a response for each.
• If the question cannot be parsed based on a context-free grammatical structure, match the question with existing questions in the database.
• If a word in the question is not in the database, consider the remaining words in the question.
• If a word in the question is not in the database, replace the word with a word in the database having a minimal number of different characters.
• The grammatical structure of the question can include an interrogative pronoun, an auxiliary verb, and a noun-phrase.
• If the question does not include a non-essential grammatical component, ignore it in the analysis.
• The question can include an interrogative pronoun, and the database can include a topic-related table for responding to the question. The question can include a non-auxiliary verb, and the database can include an event table for responding to the question. The question can include an adjective, which can be mapped to an algorithm for responding to the question. The question can include a preposition, which (a) can determine the sequence of processing other grammatical components in the question, or (b) can be mapped to an algorithm for responding to the question.
• Input can be through a voice recognition system.
• Output can be through a voice synthesizer.

Patent Number: 5,884,302 [View Patent PDF]
Date of Patent: Mar. 16, 1999

Patent Number: 5,934,910 [View Patent PDF]
Date of Patent: August 10, 1999