Joachim Van den Bogaert will be presenting MT technology innovations from CrossLang during the workshop "Web Services and Processing Pipelines in HLT: Tool Evaluation, LR Production and Validation" at LREC 2010.
The seventh international conference on Language Resources and Evaluation (LREC) is organised by ELRA in cooperation with a wide range of international associations and organisations. The conference is held in Valetta, Malta on May 19, 20 and 21 2010. LREC provides a unique forum for researchers, industrials and funding agencies from across a wide spectrum of areas to discuss problems and opportunities, find new synergies and promote initiatives for international cooperation, support investigations in language sciences, and progress in language technologies and development of corresponding products, services and applications, and standards.
Workshop on Web Services and Processing Pipelines in HLT: Tool Evaluation, LR Production and Validation
With the emergence of large e-infrastructures and the widespread adoption of the Service Oriented Architecture (SOA) paradigm, more and more language technology is being made available through web services. Extending such services to linguistic processing pipelines, tool evaluation or LR production and validation involves considering both the methodologies and technical aspects specific to the application domains.
Distributed architectures such as web services allow communication and data exchange between applications. They are a suitable instrument for automatic - less often, for semi-automatic - tool evaluation as well as resource production processes, for both practical and conceptual reasons. At a practical level, web services support quick results, centralised data storage, remote access etc.; at a conceptual level, they allow for the combination of multiple processing components which may be located on different sites. Such processing pipelines are set up to tackle a particular analysis task. To support these, new techniques have to be developed to organise well-established practices into workflows and support the exchange of data by standards and open tool architectures.
The workshop focuses on current uses and best practices for the deployment of web services and web interfaces in the HLT domain, including processing pipelines, LR production and validation, and evaluation of tools. It highlights relevant aspects for the integration of linguistic or evaluation web services within infrastructures (e.g. authorisation and authentication, service registries) and infrastructural requirements (e.g. interface harmonisation, metadata generation). The workshop also aims at demonstrating different approaches to combining linguistic web services into a composite web service.
The expected outcome of the workshop is a comparison of the practices in architectures and processing pipelines that people build, and a discussion of the issues involved. Topics of interest include, but are not limited to:
Technical aspects: approaches, protocols, management of huge amounts of data, data structures and formats, performance, manual components (e.g. annotation or evaluation), composition and configuration, interoperability, security, monitoring and recovery strategies, standardisation of APIs, tools and frameworks supporting HLT services deployment, architectures.
Scientific aspects: influence of web services on evaluation or resource production, meta-evaluation / validation of architectures, annotation agreements, needs for tools evaluation and resource production, status of the data produced.
Commercial aspects: licensing, privacy, advertising, brokering, business possibilities, challenges, exploitation of the resulting data.