Presentation is loading. Please wait.

Presentation is loading. Please wait.

Natural Language Processing Daniele Quercia Fall, 2000.

Similar presentations


Presentation on theme: "Natural Language Processing Daniele Quercia Fall, 2000."— Presentation transcript:

1

2 Natural Language Processing Daniele Quercia Fall, 2000

3 Presentation Presentation Outline Introduction to the Natural Language Processing (NLP) Applications of NLP Discussion

4 Introduction to Natural Language Processing (NLP)

5 NLP What is Natural Language processing (NLP)? “The Natural Language Processing (NLP) is the engineering of systems that process or analyse written or spoken natural language.” NLP System (computational model) Natural Language (English) Representation of the input meaning

6 Related disciplines Linguistics Computer Science Psychology Neuroscienc e Natural Language Processing

7 Knowledge of language These language processing applications use knowledge of language Unix wc program › total number of bytes, words, lines in a text file. What is a word? A possible file: PIPPO.TXT Hallo Pippo. Ciao Pippo. Auf widersehen Pippo.

8 Knowledge of language - 2 Types of knowledge used in NLP 3 Phonetics and Phonology 3 Morphology 3 Syntax 3 Semantics 3 Pragmatics 3 Discourse  NLP can be viewed as resolving ambiguity at one of these levels. i

9 Knowledge of language - 3 Example: HAL computer system. 3capable of interacting with humans Stanley Kubrik’s film 2001: A Space Odissey Conversation between HAL and Dave

10 Knowledge of language - 4 Example(cont’d): HAL computer system. 3capable of interacting with humans Stanley Kubrik’s film 2001: A Space Odissey Dave: Open the pod bay doors, right now. HAL: I’m sorry Dave, I’m afraid I can’t do that.

11 Knowledge of language - 5 Phonetics and Phonology 3The study of linguistic sounds  How word are pronounced in colloquial speech Syntax 3The study of the structural relationship between words 3but syntax don’t tell us much about the meaning … e.g. : 3I saw the Grand Canyon flying to New York 3Time flies like an arrow

12 Knowledge of language - 6 Semantics  The study of meaning e.g. : I work for J.F. Zich and he’s a good boss which is a good company Pragmatics 3The study of how language is used to accomplish goals e.g. 3Can youn tell me what time is it?  Could I have the salt?

13 Knowledge of language - 7 Morphology 3The study of the meaningful components of words. e.g. : 3Kick, kicks, kicked, kicking 3Arm, army Discourse  The study of linguistic linguistic units larger than a single utterance

14 Applications of NLP

15 ·Text-based applications ·Dialogue based applications Speech recognition Machine translation Information retrieval

16 Applications of NLP Machine translation 3Automatic machine translation 3translates texts from one language to another Windows 98 must be delivered either without Internet Explorer or in connection with the sailor of the competitor Netscape Windows 98 deve essere distribuito senza Internet Explorer o insieme al navigatore del concorrente Netscape Important constraint The text needs to be restricted to a limited range of subjects

17 Applications of NLP Machine translation 3even the best system produces poor translation… … but Machine-assisted translation uses a computer to help a translator.

18 Applications of NLP Information retrievial 3concerns the retrieving of relevant information from databases.

19 Applications of NLP Information retrievial (cont’d) 3Inverted index 3Retrieval depends only on how often each word appears in a document 3Should we treat all words equally?

20 Applications of NLP Information retrievial (cont’d) 3Stop list: words to ignore 3Many frequent words in English are function words, which are useless.

21 Applications of NLP Speech recognition 3Speech generation is comparatively easy, but recognition is hard ! 3The acoustic signal is highly ambigous 3Disambiguation using statistics If …they buy… occurs more frequently than …they by …, it will chosen

22 Conclusions

23 Conclusions 3Computers can process natural language in a variety of interesting ways 3A computer can’t do anything close to “understanding” language … 3NLP is almost never error-free

24 A very special thanks to …


Download ppt "Natural Language Processing Daniele Quercia Fall, 2000."

Similar presentations


Ads by Google