William Byrne - US grants

IIS-9820687
Jelinek, Frederick; Khudanpur, Sanjeev & Byrne, William J.
Johns Hopkins University
$500,000 - 12 mos. in FY99

1999 Language Engineering Workshop for Students and Professionals: Integrating Research and Education (WS99)

This is a one-year standard award. Automated systems that interact with human users in spoken and written communication will greatly enhance productivity and program usability. These systems will allow friendly access to information services and are essential for people with disabilities, or for accessing databases while performing intricate tasks. Unfortunately, current technology is inadequate for the tasks at hand, and there is a need to make progress. The numbers of available personnel trained in the field must be vastly increased and solutions to long standing problems must be found. At this time, relatively few universities educate students capable of performing the required tasks. Leading professionals are scattered in various industrial, adademic, and governmental institutions, often duplicating each other's work. A six-week workshop on language engineering will be conducted in which mixed teams of leanding professionals and students would fully cooperate to advance the state of the art. The professionals would normally be university professors and industrial and government researchers working in widely dispersed locations. Graduate and undergraduate students will join these teams. The participation of undergraduate students is intended not only as an educational opportunity but also to broaden the appeal of the language engineering field amongst students considering graduate studies.

Czech belongs to the family of Slavic languages, which exhibit phenomena that will prove challenging to speech recognition. The characteristics of a high degree of inflection, a high degree of derivation, a relatively free word order, and having phones that appear in no other language make speech recognition difficult. The development of a large vocabulary speaker independent continuous speech recognizer for Czech is proposed as a collaborative effort between Johns Hopkins' Center for Language and Speech Processing and three university teams from the Czech Republic. The proposed research is attempting to take advantage of the synergistic opportunities offered by the experience and tools of Johns Hopkins researchers, by the capabilities of the Czech partners, and by the existence of extensive text resources in their possession. The research cooperation will forge international scientific ties and create new centers of language engineering excellence outside the United States.

This is a standard award. Automated systems that interact with human users in spoken and written communication will greatly enhance productivity and program usability. These systems will allow friendly access to information services and are essential for people with disabilities, or for accessing databases while performing intricate tasks. Unfortunately, current technology is inadequate for the tasks at hand, and there is a need to make progress. The number of available personnel trained in the field must be vastly increased and solutions to long standing problems must be found. At this time, relatively few universities educate students capable of performing the required tasks. Leading professionals are scattered in various industrial, academic, and governmental institutions, often duplicating each other's work. This award will allow a summer workshop on language engineering to be conducted where mixed teams of leading professionals and students can cooperate to advance the state of the art. The professionals will normally be university professors and industrial and government researchers working in widely dispersed locations. Graduate and undergraduate students will join these teams. The participation of undergraduate students is intended not only as an educational opportunity, but also to broaden the appeal of the language engineering field amongst students considering graduate studies.

This is a standard award. Automated systems that interact with human users in spoken and written communication will greatly enhance productivity and program usability. These systems will allow friendly access to information services and are essential for people with disabilities, or for accessing databases while performing intricate tasks. Unfortunately, current technology is inadequate for the tasks at hand, and there is a need to make progress. The number of available personnel trained in the field must be vastly increased and solutions to long standing problems must be found. At this time, relatively few universities educate students capable of performing the required tasks. Leading professionals are scattered in various industrial, academic, and governmental institutions, often duplicating each other's work. This award will allow a summer workshop on language engineering to be conducted where mixed teams of leading professionals and students can cooperate to advance the state of the art. The professionals will normally be university professors and industrial and government researchers working in widely dispersed locations. Graduate and undergraduate students will join these teams. The participation of undergraduate students is intended not only as an educational opportunity, but also to broaden the appeal of the language engineering field amongst students considering graduate studies.

The project proposes to advance the state of the art in automatic
speech recognition by detecting emotional and highly accented speech
and differences based on age and gender, and then optimizing the
acoustic model for those conditions. It will apply long-term
adaptation techniques to improve robustness, and will implement
an innovative second-pass decoding technique to improve accuracy
by using side information such as thesaurus terms and
human-prepared summaries.

The techniques to be developed will dramatically improve the efficiency
of professional catalogers, leveraging automatic segmentation to
suggest topic boundaries in interviews, using domain-tuned
classification algorithms to recommend thesaurus terms, and
providing automated tools to support generation of event timelines.
Volunteers will help assign metadata and provide transcripts.
Efforts will be made to automate transferring capabilities developed
originally for English to other languages. Access to multilingual
materials will be done by combining knowledge-based and corpus-based
techniques to extend existing thesauri to new languages and by supporting
cross-language searching of manually prepared segment-level
summaries and automatic speech recognition transcripts.

Each component will be evaluated and user studies done to measure
the overall impact on support for cataloging, search and exploration.
This will produce significant impact, both through improved access
to our cultural heritage and through the application of the techniques
to other important problems.

The collection of spoken material used in this project will be the 116,000
hours held by the Survivors of the Shoah Visual History Foundation, a
set of already digitized video recordings of great historical importance.

We propose to organize yearly, intensive six-week research workshops
at Johns Hopkins University, focusing on Language Engineering,
including the mutually related areas of automatic speech recognition
and synthesis, natural language processing, machine translation,
information extraction and summarization. Applications of language
engineering techniques to other domains such as language instruction
or bioinformatics will also be appropriate workshop topics. These
workshops will bring together teams of leading professionals and
graduate and undergraduate students in a cooperative effort to
advance the state of the art. The first goal of the proposed workshop
series is to establish new research directions in Language Engineering.
The second goal is to attract students to the field and to offer
them an intensive hands-on mentored research education. The third
is to engender extensive cross-fertilization between researchers
distributed across industrial, academic and government institutions,
forging intensive links in the 6-week summer period and offering
a shared research environment for follow-on collaborative work.
To ensure that the projects address current problems in the state
of the art, each year an open call for workshop project proposals
will be issued to researchers in the worldwide language engineering
community. The proposals received will be evaluated competitively
at planning meetings held each year. The meetings will draw together
project proponents, government representatives, and experts from
related fields to assess the viability and promise of the proposals
and to identify three candidate projects for the summer workshop.
Throughout this process of soliciting proposals and recruiting
the personnel to carry them out, we will make a diligent attempt
to include researchers from underrepresented institutions and
communities.

The goal of this project is to enable information retrieval systems to handle speech-based information by considerably enhancing speech recognition technology. The PI's approach is to align collections of stories from text and speech sources in multiple languages, and to develop methods which exploit the resulting parallelism. The PI will improve acoustic and language models, and will develop iterative recognition methods to facilitate capture of cross-lingual information resident in text and speech. Voice-of-America broadcasts in English, Greek, and a third language to be determined, will provide the data source for this research.