XIG

The Interchange Format (IF) is based on a predefined set of Dialogue Acts (DAs) with attributes. We developed a generator that takes IF representations in input and feeds in a speech synthesizer with full/part of sentences. The speech-to-speech translation scenario imposes a number of constraints which add to those common to a generation task. First, the input representation is partially underspecified and its full interpretation relies on implicit domain knowledge. Second, the output must be in spoken language, which is characterized by frequent use of idiomatic forms, fragmentary phrases, etc. Third, the application scenario imposes a strong requirement for time efficiency and robustness. As an answer to these constraints we developed a generator based on Mixed Representations, a solution that allowed us to achieve high efficiency. The system is implemented in Sicstus Prolog running under Unix.

Try XIG ON-LINE.

People Involved

• Emanuele Pianta
• Nadia Mana
• Lucia Tovena

Publications

• Robust Generation for Speech Translation. Emanuele Pianta. Slide presentation at Recent Advances in Speech Translation Systems, ESSLLI-2002 Tutorial Course, Trento, August 12-16, 2002, [ get .ps.gz]
• XIG: Generating from Interchange Format using Mixed Representations. Emanuele Pianta and Lucia M. Tovena. In Proceedings of the sixth AIIA Conference, Bologna I, September 1999 [ get .ps]
• Comparing methodologies for evaluating the generator in a speech-to-speech translation system. Fabio Pianesi, Emanuele Pianta and Lucia M. Tovena. In Proceedings of the European Workshop on Natural Language Generation (EWNLG'99), Toulouse F, 1999, 145--154
• Mixing representation levels: The hybrid approach to automatic text generation. Emanuele Pianta and Lucia M. Tovena. In Proceedings of the Artificial Intelligence and Simulation of Behaviour Workshop on ``Reference Architectures and Data Standards for NLP'', Edinburgh Scotland, 1999, 8-13. [get .ps]
• Generating felicitous sentences from underspecified semantic representations. Lucia M. Tovena and Emanuele Pianta. In Proceedings of the 3rd International Workshop on Computational Semantics 1999, Tilburg NL , 410-412. [get .ps]
• Using the Interchange Format for encoding spoken dialogues. Fabio Pianesi and Lucia M. Tovena. In Proceedings of AMTA SIG-IL, Second Workshop on Interlinguas and Interlingual Approaches, Pennsylvania USA, 1998.
• A user's point of view on the interlingua IF. Lucia Tovena, ITC-irst Technical Report 9809-01, 1998. [get .ps]

Maintainer: