CLARIN-NL - speech processing

TTNWW

root — Sun, 15 Jun 2014 13:01:54 +0000

TTNWW - TST Tools voor het Nederlands als Webservices in een Workflow

Summary

TTNWW integrates and makes available existing Language Technology (LT) software components for the Dutch language that have been developed in the STEVIN and CGN projects. The LT components (for text and speech) are made available as web-services in a simplified workflow system that enables researchers without much technical background to use standard LT workflow recipes.

Background

The web services are available in two separate domains: "Text" and "Speech" processing. For "Text", workflows for the following functionality is offered by TTNWW:

Orthographic Normalisation using TICCLops (version CLARIN-NL 1.0)
Part of Speech Tagging, Lemmatisation, Chunking, limited Multiword Unit Recognition, and Grammatical Relation Assignment by Frog (Version 012.012)
Syntactic Parsing (including grammatical relation assignment, limited named entity recognition, and limited multiword unit recognition) by the Alpino Parser (version 1.3)
Semantic Annotation
Named Entity Recognition
Co-reference Assignment

For "Speech", the following workflows are offered:

Automatic Transcription of speech files using a Netherlands Dutch acoustic model
Automatic Transcription of speech files using a Flemish Dutch acoustic model
Conversion of the input speech file to the required sampling rate, followed by automatic transcription

The TTNWW services have been created in a Dutch and Flemish collaboration project building on the results of past Dutch and Flemish projects. The web services are partly deployed in the SURF-SARA BiG-Grid cloud or at CLARIN centres in the Netherlands and at CLARIN VL University partners.

The architecture of the TTNWW portal consists out of several components and follows the principles of Service Oriented Architecture (SOA). The TTNWW GUI front-end is a Flex module that communicates with the TTNWW web-application which keeps track of the different sessions and knows which LT recipes are available. TTNWW communicates assigments (workflow specifications) to the WorkflowService that evaluates the requested workflow and requests the DeploymentSevice to start the required LT web-services. After initialization of the LT web-services, the workflow specification is sent to the Taverna Server, that takes further care of the workflow.

To facilitate the process of wrapping applications that were originally designed as standalone applications into web services, the CLAM (Computational Linguistics Application Mediator) wrapper software allows for easy and transparent transformation of applications into RESTful web services. The CLAM software has extensively been used in the TTNWW project for both text and speech processing tools. With the exception of Alpino and MBSRL all web services work operate on CLAM wrappers.

Given the number of web services involved in the TTNWW project and possibilities offered by the cloud environment the preferred method of delivering the web service installations was delivery of complete virtual machine images by the LT providers. These could be directly uploaded into the cloud environment and thus relieving the CLARIN centres nd LT providers from the original foreseen task of running the webservices themselves. A potential advantage of this method, that has not been exploited in the project yet, is that these images may be also be delivered directly to the end user so these can be run in a local configuration using virtualization software such as VMWare of VirtualBox.

The workflow engine used in the project was Taverna. But build on top of this was a a number of selectable task recipes, following a task oriented approach in line with the premises that users with no or little technical expertise should be able to use the system. In this context, tasks are understood in terms of end results of processes such as semantic role labelling, pos tagging or syntactic analysis and ready-made workflows are constructed that can be readily used by the end user. Contacts

Project leader: Marc Kemps-Snijders (NL), Ineke Schuurman (VL)  
CLARIN center: Meertens Institute (portal) and others (web-services).
Help contact : n.a.

Links

Web-sites: n.a.
User scenario's (screencasts, screenshots): n.a.
Manual: http://yago.meertens.knaw.nl/apache/TTNWW/assets/TTNWW.pdf
Tool/Service link: http://yago.meertens.knaw.nl/apache/TTNWW/
Publications: n.a.

Research domain

Communication & Media Studies

History

Oral History

Linguistics

Morpho-syntax

Orthography

Discourse

Semantics

Syntax

Language

Dutch

Country

Netherlands

DCS

root — Mon, 09 Jun 2014 09:40:50 +0000

DCS - Digital Curation Service

Summary

The Digital Curation Service was a CLARIN task force responsible for the curation of several language resources that were not available on-line and/or endangered without a proper documentation context (adding CMDI metadata etc.). The DCS has curated these data and made them available by depositing them at different CLARIN centra.

Background

In October 2011 the Data Curation Service (DCS) was established. The DCS was supported by CLARIN-NL until 1 January 2014. The DCS was charged with:

the curation of resources, especially those presently held by individual researchers or research groups;
assisting in the curation efforts of CLARIN centres (if and when such is desired);
advising researchers who wanted to undertake the curation of their resources themselves.

The curation of resources held by individual researchers or research groups comprised the core of the work undertaken by the DCS. and the project was run from the Centre for Language and Speech Technology (CLST) at Radboud University Nijmegen and led by Nelleke Oostdijk (head) and Henk van den Heuvel (deputy head).

The DCS was involved in the curation of the following resources:
Completed:

LESLLA (delivered in August 2013)
- CLARIN centre: MPI for Psycholinguistics
- link: http://hdl.handle.net/1839/00-37EBCC6D-04A5-4598-88E2-E0F390D5FCE1@forma...@view
IPNV Interviews with veterans (delivered in May 2013)
- CLARIN centre: DANS
- link: https://easy.dans.knaw.nl/ui/?wicket:bookmarkablePage=:nl.knaw.dans.easy...
Woordenboek Gelderse Dialecten, Rivierengebied (delivered in June 2013)
- CLARIN centre: Meertens Institute
- link: ?
Woordenboek Gelderse Dialecten, Veluwe (part) (delivered in June 2013)
- CLARIN centre: Meertens Institute
- link: ?
Curation organisation names for OpenSkos (delivered in June 2013; finalised in August 2013)
Six dialect dictionaries from Brabant (delivered in October 2013)
- CLARIN centre: Meertens Institute
- link: ?
WLD and WBD part III (delivered in December 2013)
- CLARIN centre: Meertens Institute
- link: ?
DBD/TCULT (delivered in February 2014)
- CLARIN centre: MPI for Psycholinguistics
- link: https://corpus1.mpi.nl/ds/asv/?openpath=node:84720

Unfinished (and being considered by CLARIN centres for adoption):

Roots of Etnolect (collection Linda van Meel)
Traces of Contact (collection Kofi Yakbo)

Contacts

Project leader: Nelleke Oostdijk
CLARIN center: several
Help contact : https://www.ru.nl/letteren/datacuratieservice/informatie/contact/

Links

Web-sites: http://www.ru.nl/letteren/datacuratieservice/
User scenario's (screencasts, screenshots): n.a.
Manual: n.a.
Tool/Service link: n.a.
Publications:
- Oostdijk, N. & H. van den Heuvel. 2012. Introducing the CLARIN - NL Data Curation Service. In Proceedings of the Workshop Challenges in the management of large corpora. LREC2012, Istanbul, 22 May 2012. http://www.lrec-conf.org/proceedings/lrec2012/index.html . Retrieval date: 20 March 2014.
- Heuvel, H. van den, Oostdijk, N.H.J., Sanders, E.P. & Lint, V. de (2014). Data curations by the Dutch Data Curation Service. In Proceedings CLARIN Annual Conference (pp. online). http://www.clarin.eu/sites/default/files/cac2014_submission_15_0.pdf
- Oostdijk, N., Heuvel, H. van den, Treurniet, M. (2013): The CLARIN-NL Data Curation Service. Bringing data to the foreground. International Journal of Digital Curation, Vol 8, No 2 (2013), pp. 134-145.

Resource tags

service

speech processing

text processing

Country

Netherlands

TQE

root — Sun, 20 Apr 2014 13:45:49 +0000

TQE: Transcription Quality Evaluation

Summary

The Transcription Quality Evaluation tool can be used to check the quality of phonetic transcription. The only thing the researcher has to do is upload pairs of files consisting of an audio file and a transcription file. After uploading he’ll receive an e-mail with the matching output.

Background

The Transcription Quality Evaluation (TQE) tool is an instrument that automatically evaluates the quality of phonetic transcriptions. The application makes it possible to upload pairs of files consisting of an audio file and a transcription file and process them as follows: the audio signal and the phonetic transcription are aligned, segment boundaries are derived for each phone, and for each segment-phone combination it is determined how well they fit together, i.e. for each phone a TQE measure (a confidence measure) is determined, a number ranging from 0-100%, indicating how good the fit is, i.e. the quality of the phone transcription. The higher the number, the better the fit. The output of the TQE tool consists of a TQE measure and the segment boundaries for each phone in the corpus.
The TQE tool thus makes it possible to find (sequences of) segments for which the match of the phone symbols with the audio signal is not optimal, in other words, the TQE tool can be used to check the quality of phonetic transcriptions. This can be useful for validating (manual) phonetic transcriptions, but also to compare and select (‘competing’) transcriptions, e.g. to study pronunciation variation. The TQE tool can thus be usefully applied in all research – in various (sub-) fields of humanities and language and speech technology (L&ST) – in which audio and phonetic transcriptions are involved.

Contacts

Project leader:  Dr. H. Strik (Radboud University)
CLARIN center: Max Planck Institute for Psycholinguistics (MPI)
Help contact : http://vps8639.xlshosting.net/TQE/?p=contact

Links

Web-sites: http://vps8639.xlshosting.net/TQE/
User scenario's (screencasts, screenshots): n.a.
Manual: http://vps8639.xlshosting.net/TQE/?p=manual
Tool/Service link: http://hdl.handle.net/1839/00-SERV-0000-0000-0005-6
Publications: n.a.

Research domain

Speech recognition

Linguistics

Phonetics

Language

Language independent

Resource tags

service

speech processing

web-application

Tool task

evaluation

annotation

CLARIN centre

MPI for Psycholinguistics

Country

Netherlands

AAM-LR

root — Mon, 14 Apr 2014 18:05:02 +0000

AAM-LR: Automatic Annotation of Multi-modal Language Resources

Summary
The AAM-LR web service helps researchers to annotate audio and video recordings. At the top level the service marks the time intervals at which specific persons in the recording are speaking. In addition, the service provides a global phonetic annotation, using language independent phone models and phonetic features. Speech is separated from speaker noises such as laughing. The output of the web service is fed into the ELAN/ANNEX editor, to facilitate further manual annotation. The annotations conform to ISOCat and potential new categories were added to ISOCat.

Background
Analyzing large amounts of field work recordings is prohibitively time consuming and expensive. As a consequence, there is a rapidly growing pile of data that does exist, but that is practically impossible to use. It is estimated that by the year 2020 more than 90% of the field work recordings will be sitting unused in archives, unless automatic tools for annotating the data can be made available. Given the cost of the field work and making the recordings in the first place, the lack of efficient tools for processing the data is especially awkward. While the AAM-LR tool will not provide fully automatic analysis of the field work corpora, they will certainly reduce the time needed for making linguistic analysis by several orders of magnitude, thereby bringing effective usage of the data within the realm of what is feasible. Thus, the AAM-LR tools will not help with respect to a specific research question, but they will improve the working environment for linguists working in areas of language contact, language typology, language and cognition and many other sub-disciplines. The AAM-LR tools will also make fieldwork recordings accessible for other types of humanities research. [site:mail]

Contacts

Project leader:  Em. prof. dr. Lou Boves (Radboud University)
CLARIN center : Max Planck Institute for Psycholinguistics
Help contact : http://tla.mpi.nl/contact/

Web-site s: n.a.
User scenario's: n.a.
Manual : n.a.
Tool/Service link: 
- http://catalog.clarin.eu/aamlr/
- http://hdl.handle.net/1839/00-SERV-0000-0000-004F-F
Publications: Prawiro-Atmodjo, P., L. ten Bosch, 2010. Report on the AAM-LR project. RU NL-CLARIN report, RU Nijmegen. (Includes AAM-LR ISOCAT extension proposal.)