	home \| site map \| contact us

	About Us

	Consortium

	Technology

	Projects • Classroom Testing • Net4Voice • Research & Development • Corporate Access • Community Access • RealTime Interaction

	Resource Centre

	News & Events

Projects

RESEARCH AND DEVELOPMENT

OVERVIEW

Research and development continues with ViaScribe and new technology applications. Projects are developed in collaboration with IBM Research and other partners, yet reflect individual consortium member interests. All projects are pursued within a technology transfer model that synergizes developments across the group. University classrooms around the world are the living laboratory for ViaScribe, providing user feedback, usability testing, and the rationale for development priorities.

Examples of recent improvements to IBM ViaScribe are: a new Java-based GUI, integrated editing and display, PowerPoint custom applications, an improved editing GUI, pedal devices, HTML output, mobile interface, and annotations capability.

The research and development projects described on this page are all ongoing, and in many cases will improve features already in use with ViaScribe.

RESEARCH AND DEVELOPMENT PROJECTS

Interface and Functional Development
Lead: IBM Research, Saint Mary's University

Areas of investigation include:
- Improved Support for Internationalization: speaker independent dictation introduced to Japanese and UK-English (currently available in US-English, Chinese, French, and Italian).
- Easier Installation: language models bundled separately to vastly reduce size of setup program.
- Improved User Interface: new wizard-based training tools, reorganized menus.
- Simpler File Naming Convention
- Customizable Multimedia Transcripts: ability to change the way ViaScribe saves interactive content.
- Improved Development Strategy: separation of ViaScribe development into two streams:"alpha" (experimental) and "beta" (release-candidate).
- Pause Button: halt transcription without creating a new session.
- Engine Upgrade: integration of new research-grade SR engine in two phases
  
  i) "offline" transcription (transcription from file)
  
  ii) live transcription from microphone.
- New Multimedia Format: Flash-based for cross-browser / cross-platform support.
- Support for Meta Data: semantic markup of transcribed text through "Annotations".
- Improved Training Tool: grid interface to make third-party training of profiles easier and more efficient.
- ViaScribe API: access to ViaScribe functionality for extension developers

Top of Page

Real Time Editing and Annotation
Lead: University of Southampton

It is possible to edit errors in synchronised speech and text to insert, delete or amend the text with the timings being automatically adjusted. For example, an 'editor' correcting 15 words per minute would improve the accuracy of the transcribed text from 80% to 90% for a speaker talking at 150 words per minute. Not all errors are equally important, and so the editor can use their initiative to prioritise those that most affect readability and understanding. Methods to enable real time editing to occur are being investigated (e.g. highlighting errors or lines containing errors and displaying corrections).

Since it would take students a long time to read through a verbatim transcript after a lecture and summarise it for future use, it would be valuable for students to be able to create an annotated summary for themselves in real time through highlighting, selecting and saving key sections of the transcribed text and adding their own words time linked with the synchronised transcript. Methods to enable this to occur are also being investigated.
Automatic Transcription, Summarization, Retrieval
Lead: MIT Computer Science and Artificial Intelligence Laboratory (CSAIL)

The Spoken Language Systems group in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) conducts research and develops technology that will enable humans to interact with computers using natural spoken language. One aspect of the research mission focuses on the uses of human language technology in support of education. One of the current research activities in this area is focusing on the automatic transcription, annotation, and summarization of spontaneous spoken lectures.

The goal is to create lecture processing tools that can be easily integrated into various need-dependent frameworks. Usability testing involves integrating the tools into MIT Open Courseware and as well as Liberated Learning initiatives.

This evaluation framework is particularly valuable for applications such as summarization, where researches have little experience with task-based evaluation. A large corpus of task-based utility assessments will facilitate the development of automatic evaluation measures that do not directly involve user judgment. In addition, evaluation in the context of the two projects will allow us to consider the impact of different user goals on their assessment of new technology.
Intelligent Search and Retrieval: Note Finder
Lead: Saint Mary's University

The Mathematics and Computing Science department at Saint Mary's University has developed Note Finder, an indexing and phrase table system used in conjunction with Liberated Learning's speech recognition generated notes.

Note Finder provides students with improved methods to use the online contents by:

automatically indexing uploaded course notes

allowing students to search for specific words or phrases

using phrase and proximity lists for more effective searching and retrieval

using stemmed words based on the Paice/Husk stemming rules

The system includes a search engine and an index table for the entire course, taking into account that content is in the form of spoken language. It flags words that go together to form phrases and these words are grouped based on their relative distance. The Mathematics and Computing Science department currently hosts the facility as a service, providing technical support and statistical information during its initial developmental stage.

Future efforts will focus on the inclusion of various formats such as Microsoft Word, PowerPoint, Excel, PDF, and multimedia notes created by ViaScribe.