Challenges and Goals
PANACEA´s objective is to join together a number of advanced interoperable tools to build a factory of LRs, a production line that automates all of the stages involved in the acquisition, production, updating and maintenance of LRs required by MT and other Language Technologies.
Cost and time reduction by automation is the only way to ensure the continuous supply of LRs needed by the LT industry to solve current and future needs, as well as, problems such as MT systems covering all languages and domains in the time required by the market.
The factory will be built as a Web Service-based platform for easy integration of the latest technological components for:
- Monolingual and Parallel Text Acquisition and Pre-processing
- Sentential and Sub-Sentential Alignment
- Bilingual Dictionary and Transfer Grammar Production
- Lexical Information Acquisition for rich information dictionary production.
Our understanding is PANACEA addresses key challenges for the future of MT and other LT applications. A quantitative objective that gives a clear idea of the impact of PANACEA´s results is that it's aim to reduce the actual cost (both in human resources and time) of manually building a domain-tuned language resource required by MT providers by more than, on average, 25%. These improvements will be measured by an evaluation carried out in an industrial scenario for a Rule-Based MT system and in a pre-marketing scenario for a Statistical MT system.
