1 / 14

Language Technology – as a field of specialization

Language Technology – as a field of specialization. Udaya Narayana Singh CIIL UGC-ASC Refresher Course In Applied Linguistics. What is Language Technology?.

pbreton
Download Presentation

Language Technology – as a field of specialization

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Language Technology –as a field of specialization Udaya Narayana Singh CIIL UGC-ASC Refresher Course In Applied Linguistics

  2. What is Language Technology? • Language Technology is the term we use to describe a range of computational techniques designed to processreal human language, whether that language presents itself in spoken or textual form. • To a large extent, language technology encompasses those techniques and applications that are more traditionally seen as being the domain of natural language processing (NLP).

  3. Construction of speech recognition systems Machine translation systems, or at least MAT systems Grammar checking systems Natural language database interfaces Research carried out in laboratory conditions Actual context of use of such techniques are abstracted away here Deals with language as it appears in everyday use – in paper documents, telephone lines, or on the World Wide Web Typical concerns

  4. The Focus • Create products with practical applications • Learn more about how language actually works in real life contexts by building computational models of language use or language applications. • Decide as to how to deal with noisy or errorful data, or how to handle punctuation and formatting • Enable people to communicate with machines using natural communication skills • R&D involves coding, recognition, interpretation, translation, and generation of language

  5. Are you on the side of LT? • Advances in computer science necessary to create the architectures and platforms to represent this knowledge. • Inter-disciplinary collaboration is needed to create multimodal and multimedia systems that combine speech, facial cues and gestures - to improve lang. understanding and to produce more natural and intelligible speech by animated characters. • As the importance of interactive networks increases in commerce and daily life, those who do not have access to computers or the skills to use them are further handicapped from becoming productive members of society.

  6. The 1993 Human Language Technology Workshop at the Merril Lynch Conference Center in New Jersey made Oscar Garcia & Antonio Zampolli come up with a comprehensive survey of LT • .

  7. Requires expertise in areas of linguistics, psychology, computer science & engg. Machines to interact with people naturally requires a deep understanding of the acoustic and symbolic structure of language Also requires understanding of communication mechanisms and strategies Given the remarkable ability of people to converse under adverse conditions, such as noisy social gatherings or band-limited communication channels, advances in signal processing are essential to produce robust systems (the domain of electrical engineering) Multi-disciplinarity

  8. Goal: Inverse formatting or completion of the Gutenberg loop, i.e. a scanned printed doc is translated back into a document description language from which it could be accurately reprinted if desired. So far worked for technical papers, business letters and chemical structure diagrams. PROBLEMS:Some OCRs provide limited inverse formatting, producing codes for elementary structures such as paragraphs, columns, and tables, but current OCRs cannot still encode, halftones and line drawings. Another application: Style identification which assists users to locate and recognize the title, author, and abstract of each article, and to extract keywords. But it is still limited to technical articles. DIA Application on Text Docs

  9. 2.Second challenge in Application : Forms • Forms are the printed counterparts of relations in a data base – typically, an n-tuple of data items each represented as an ordered pair (item name, item value). OCR is used to recognize the item value. • Capability for locating items on a form, establishing their name class, and encoding the accompanying data values has applications in business and government. • Form DIAs work in a single enterprise and single application where they are highly repetitive in structure (such as in Income-Tax returns or Sale-tax forms) from one example to the next provides a good application. This requires extraction of data from a large variety of forms, which can be done.

  10. Follows a well-defined logical format (but a highly variable physical layout), and has a high degree of contextual constraint on the symbolic data – hence a high accuracy rate. Contextual rules can modify OCR results to force agreement of city names and postal codes, or numeric amounts on checks. Postal machines can now read the complete address fields. They also arrange mailpieces into delivery order for the route of individual postmen. Line Drawings Activity centers around CADCAM systems, as it shows that analysis of input of integrated circuit can help predict new circuits efficiently. The problem is that manual is still better than direct input at a terminal, due to the small screen sizes. Residual Applications POSTAL APPLIC:

  11. Future Needs: Where Linguists could come in • DIA systems provide fast storage, recall and distribution of • Helps with indexing & partition the image into subregions of interest for convenient access by users. • Now being extended to the creation of electronic libraries. • There is a strong need to incorporate contexts. • 35 years of R&D have still not been able to produce OCR based on shape that has the accuracy of human vision. • Linguistic analysis may not simply be a postprocessing stage in future document analysis systems. • Future Need: Many scanned document data bases - each representative of text, engineering drawings, addresses, forms, handwritten manuscripts, etc. documents. Currently there are only text-oriented documents.

  12. Natural Language Understanding (NLU) • Spoken language interfaces to computers fascinated all for over five decades. • To converse with a machine - the ultimate challenge to LT. • It is fast becoming a necessity. Such interactive networks today are limited to people who can read/use internet - resulting in further stratification of society . • NLU will provide easy access to a wealth of information and services & will fundamentally affect our daily affairs.

  13. A Model: • Spoken input to PCs embodies many different technologies/applicationsas shown here. • The biggest challenge was the conversion of an acoustic signal to a stream of words • Current applications: voice dialing, call routing,, simple data entry, and preparation of structured documents (say, a radiology report). • Speaker recognition can involve identifying a specific speaker out of a known population, which has forensic implications

More Related