| 1. |
- Becket, Ralph, et al.
(författare)
-
Spoken Language Translator: Phase Two Report
- 1997
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- Spoken Language Translator (SLT) is a project whose long-term goal is the construction of practically useful systems capable of translating human speech from one language into another. The current SLT prototype, described in detail in this report, is capable of speech-to-speech translation between English and Swedish in either direction within the domain of airline flight inquiries, using a vocabulary of about 1500 words. Translation from English and Swedish into French is also possible, with slightly poorer performance.A good English-language speech recognizer existed before the start of the project, and has since been improved in several ways. During the project, we have constructed a Swedish-language recognizer, arguably the best system of its kind so far built. This has involved among other things collection of a large amount of Swedish training data. The recognizer is essentially domain-independent, but has been tuned to give high performance in the air travel inquiry domain.The main version of the Swedish recognizer is trained on the Stockholm dialect of Swedish, and achieves near-real-time performance with a word error rate of about 7%. Techniques developed partly under this project make it possible to port the recognizer to other Swedish dialects using only modest quantities of training data.On the language-processing side, we had at the start of the project a substantial domain-independent language-processing system for English, a preliminary Swedish version, and a sketchy set of rules to permit English to Swedish translation. We now have good versions of the language-processing system for English, Swedish and French, and fair to good support for translation in five of the six possible language- pairs. Translation is carried out using a novel robust architecture developed under the project. In essence, this translates as much of the input utterance as possible using a sophisticated grammar-based method, and then employs a much simpler set of word- to-word translation rules to fill in the gaps.The language-processing modules are all generic in nature, are based on large, linguistically motivated grammars, and can fairly easily be tuned to give good performance in new domains. Much of the work involved in the domain adaptation process can be carried out by non-experts using tools developed under the project.Formal comparisons are problematic, in view of the different domains and languages used and the lack of accepted evaluation criteria. None the less, the evidence at our disposal suggests that the current SLT prototype is no worse than the German Verbmobil demonstrator, in spite of a difference in project budget of more than an order of magnitude.
|
|
| 2. |
- Björk, Staffan, et al.
(författare)
-
WEST: A Web Browser for Small Terminals
- 1999. - 1
-
Konferensbidrag (refereegranskat)abstract
- We describe WEST, a WEb browser for Small Terminals, that aims to solve some of the problems associated with accessing web pages on hand-held devices. Through a novel combination of text reduction and focus+context visualization, users can access web pages from a very limited display environment, since the system will provide an overview of the contents of a web page even when it is too large to be displayed in its entirety. To make maximum use of the limited resources available on a typical hand-held terminal, much of the most demanding work is done by a proxy server, allowing the terminal to concentrate on the task of providing responsive user interaction. The system makes use of some interaction concepts reminiscent of those defined in the Wireless Application Protocol (WAP), making it possible to utilize the techniques described here for WAP-compliant devices and services that may become available in the near future.
|
|
| 3. |
- Bretan, Ivan, et al.
(författare)
-
Approaches to gathering realistic training data for speech translation systems
- 1996
-
Ingår i: Proceedings of Third IEEE Workshop on Interactive Voice Technology for Telecommunications Applications, 1996. - : Institute of Electrical and Electronics Engineers (IEEE). - 0780332385 ; , s. 97-100
-
Konferensbidrag (refereegranskat)abstract
- The Spoken Language Translator (SLT) is a multi-lingual speech-to-speech translation prototype supporting English, Swedish and French within the air traffic information system (ATIS) domain. The design of SLT is characterized by a strongly corpus-driven approach, which accentuates the need for cost-efficient collection procedures to obtain training data. This paper discusses various approaches to the data collection issue pursued within a speech translation framework. Original American English speech and language data have been collected using traditional Wizard-of-Oz (WOZ) techniques, a relatively costly procedure yielding high-quality results. The resulting corpus has been translated textually into Swedish by a large number of native speakers (427) and used as prompts for training the target language speech model. This ￜbudgetￜ collection method is compared to the accepted method, i.e., gathering data by means of a full-blown WOZ simulation. The results indicate that although translation in this case proved economical and produced considerable data, the method is not sensitive to certain features typical of spoken language, for which WOZ is superior
|
|
| 4. |
- Bretan, Ivan, et al.
(författare)
-
Corpora and Data Collection
- 2000
-
Ingår i: The Spoken Language Translator. - Cambridge : Cambridge University Press. - 0521770777 - 9780521770774 ; , s. 131-144
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- This book presents a detailed description of Spoken Language Translator (SLT), one of the first major projects in the area of automatic speech translation. The SLT system can translate between English, French, and Swedish in the domain of air travel planning, using a vocabulary of about 1500 words, and with an accuracy of about 75%. The greater part of the book describes the language processing components, which are largely built on top of the SRI Core Language Engine, using a combination of general grammars and techniques that allow them to be rapidly customized to specific domains. Speech recognition is based on Hidden Markov Mode technology, and uses versions of the SRI DECIPHER system. This account of the Spoken Language Translator should be an essential resource both for those who wish to know what is achievable in spoken-language translation today, and for those who wish to understand how to achieve it.
|
|
| 5. |
- Bretan, Ivan
(författare)
-
Natural Language in Model World Interfaces
- 1995. - 1
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- An essay on rationale and methodology of combining natural language and other means of interaction in multimodal interfaces.
|
|
| 6. |
- Bretan, Ivan, et al.
(författare)
-
Synergy Effects in Natural Language-Based Multimodal Interaction
- 1994. - 1
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- We discuss the synergetic effects that can be obtained in an integrated multimodal interface framework comprising on one hand a visual language-based modality and on the other natural language analysis and generation components. Besides a visual language with high expressive power, the framework includes a cross-modal translation mechanism which enables mutual illumination of interface language syntax and semantics. Special attention has been payed to how to address problems with robustness and pragmatics through unconventional methods which aim to enable user control of the discourse management process.
|
|
| 7. |
- Bretan, Ivan, et al.
(författare)
-
Synergy Effects in Natural Language-Based Multimodal Interaction
- 1993. - 1
-
Ingår i: Proceedings of the Workshop on Multimodal Human-Computer Interaction.
-
Konferensbidrag (refereegranskat)abstract
- A scheme for integration of a natural language interface into a multimodal environment is presented with emphasis on the synergetic results that can be achieved, which are argued to be: 1)Complementary expressiveness; 2) Mutual illumination of language syntax and semantics; 3) Robust pragmatics and graceful recovery from failed natural language analyses through the reification of discourse objects to enable user control of discourse management.
|
|
| 8. |
- Bretan, Ivan, et al.
(författare)
-
Worlds Without Words
- 1994. - 3
-
Ingår i: Proceedings of the Workshop on Distributed Virtual Environments. ; , s. 211-216
-
Konferensbidrag (refereegranskat)abstract
- This paper points out a deficiency in virtual worlds as they are envisioned today. Several of the more important aspects of human communicative behavior are missing, since there is no way to accommodate natural linguistic interaction with and through the interface. We point out some explicit cases where natural language interaction would be useful in a VR framework, and identify some areas where natural language interpretation actually might benefit from being situated in a highly formalized environment.
|
|
| 9. |
- Dewe, Johan, et al.
(författare)
-
Assembling a Balanced Corpus from the Internet
- 1998. - 5
-
Konferensbidrag (refereegranskat)abstract
- For empirically oriented textual research it is crucial to have materials available for extraction of statistics, training probabilistic algorithms, and testing hypotheses about language and language processing in general. In recent years, the awareness that text is not just text, but that texts comes in several forms, has spread from more theoretical and literary subfields of linguistics to the more practically oriented information retrieval and natural language processing fields. As a consequence, several test collections available for research explicitly attempt to cover many or most well-established textual genres, or functional styles in well-balanced proportions (Francis and Kucera, 1982; Källgren, 1990). The creation of such a collection is a complex matter in several respects. Our reseach area is to build retrieval tools for the Internet, and thus, for our purposes, the choice of genres to include is one of the more central problems: there is no well-established genre palette for Internet materials. To find materials to experiment with, we need to create them in a form suitable for our purposes. This is a double edged problem, involving both vaguely expressed user expectations and establishing categories using large numbers of features which taken singly have low predictive and explanatory power. This paper gives an outline of the methodology we use for determining which genres to include.
|
|
| 10. |
- Eklund, Robert, 1962-, et al.
(författare)
-
Rational Re-Use of Linguistic Data
- 2000
-
Ingår i: The Spoken Language Translator. - Cambridge : Cambridge University Press. - 0521770777 - 9780521770774 ; , s. 212-228
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- This book presents a detailed description of Spoken Language Translator (SLT), one of the first major projects in the area of automatic speech translation. The SLT system can translate between English, French, and Swedish in the domain of air travel planning, using a vocabulary of about 1500 words, and with an accuracy of about 75%. The greater part of the book describes the language processing components, which are largely built on top of the SRI Core Language Engine, using a combination of general grammars and techniques that allow them to be rapidly customized to specific domains. Speech recognition is based on Hidden Markov Mode technology, and uses versions of the SRI DECIPHER system. This account of the Spoken Language Translator should be an essential resource both for those who wish to know what is achievable in spoken-language translation today, and for those who wish to understand how to achieve it.
|
|
