Congratulations to Prof. Anna Korhonen and her team at the Cambridge Language Technology Laboratory (LTL) on being awarded a European Research Council (ERC) Proof of Concept grant of €150,000 for their project, 'Enabling Multilingual Conversational AI'.
Proof of Concept grants are part of the EU's research and innovation programme, Horizon 2020, and can be used to explore the commercial or societal potential of research results – such as exploring business opportunities, preparing patent applications or verifying the practical viability of scientific concepts.
Enabling Multilingual Conversational AI will build on the work of Prof. Korhonen’s ERC-funded LEXICAL project, which has developed a novel approach to lexical acquisition in Natural Language Processing (NLP) with huge potential to enhance the viability and portability of NLP technology within and across languages.
The team, led by Prof. Korhonen and including Dr Ivan Vulić, Senior Research Associate, and Edoardo Ponti, PhD student in the LTL, will use the new ERC grant to develop the prototype for a new multilingual language understanding system that supports the porting of task-based statistical dialogue systems (SDS) to multiple languages based on guided, linguistically informed cross-lingual transfer.
Although current task-oriented conversational AI systems – such as those used in applications like Apple Siri, Amazon Echo, Google Assistant and Microsoft Cortana – are capable of supporting humans in many practical tasks, they are limited to a small number of resource-rich languages such as English and German. Extending the scope to additional and lower-resource languages could significantly improve the usability and global reach of this technology, and mitigate the effects of the digital language divide.
Anna Korhonen is a Professor of Computational Linguistics at the University of Cambridge where she co-directs the Language Technology Laboratory (LTL). Research in LTL is centred around Natural Language Processing (NLP) and how to develop, adapt and apply fundamental NLP techniques to meet the needs of intelligent applications. Based in the School of Arts and Humanities, LTL is particularly interested in human-centric NLP that draws on the understanding of human cognitive, social and creative intelligence and focuses on applications aimed at social good. Many of LTL’s projects are interdisciplinary and involve collaboration with researchers across different subject areas.