From NLG Systems Wiki

Jump to: navigation, search


[edit] Summary


Amalgam is a novel system developed in the Natural Language Processing group at Microsoft Research for sentence realization during natural language generation. Sentence realization is the process of generating (“realizing”) a fluent sentence from a semantic representation. From the outset, the goal of the Amalgam project has been to build a sentence realization system in a data-driven fashion using machine learning techniques. Amalgam accepts as input a logical form graph capturing the meaning of a sentence. Amalgam transforms the logical form into a fully articulated tree structure from which an output sentence is read. To date, we have implemented Amalgam for both German and French, with English in the works.

Example of Amalgam input


Example of Amalgam structural output



  1. Gamon, M., Ringger, E., Zhang, Z., Moore, R., & Corston-Oliver, S. (2002). Extraposition: A case study in German sentence realization. Paper presented at Proceedings of COLING 2002. Bib
  2. Corston-Oliver, S., Gamon, M., Ringger, E., & Moore, R. (2002). An overview of Amalgam: A machine-learned generation module. Paper presented at Proceedings of the International Natural Language Generation Conference, New York, USA. Bib
  3. Smets, M., Gamon, M., Corston-Oliver, S., & Ringger, E. (2003). The adaptation of a machine-learned sentence realization system to French. Paper presented at Proceedings of the 10th conference of the European Chapter of the Association for Computational Linguistics, Budapest, Hungary. Bib
Facts about AmalgamRDF feed
DescriptionAmalgam is a sentence realization component based on data-driven, machine learning techniques. Amalgam accepts as input a logical form graph.  +
Domaintechnical manuals  +
FrameworkMachine learning  +
LanguageGerman  +, French  +, and English  +
NameAmalgam  +
Started2001  +
URL  +
WorkerGamon  +, Ringger  +, Corston-Oliver  +, and Moore  +
Personal tools