System ASIUM
Version 1.0
Code ANSI C, interface in Tcl/Tk. All platforms supporting Tcl/Tk.
References [1], [2], [3]

ASIUM is a system based on clustering and cooperative methods.

ASIUM takes as input the syntactic analysis of sentences such that the phrases are identified and for each noun phrase the preposition if any and the head word IS given.

ASIUM learns an ontology for texts given as input and a set of case frames where the semantic features are filled by ontology concepts.

The method is based on conceptual and hierarchical clustering. Basic clusters are formed by words that occur with the same verb after the same preposition (or the same function, if there is no preposition). A distance is computed between all basic clusters. New clusters are formed by gathering the closest clusters. The hierachy of clusters form the ontology. Case frames are learned in parallel, so that each time a concept is formed, it fills the corresponding semantic feature in the case frame.

The graphic interface of ASIUM allows the user to inspect, validate and refine the knowledge learned at each step of the learning process.


  1. Claire Nédellec. Corpus-based learning of semantic relations by the ILP system, Asium. In James Cussens, editor, Proceedings of the 1st Workshop on Learning Language in Logic, pages 28--39, Bled, Slovenia, June 1999.

  2. Faure D. and Nédellec C. Knowledge acquisition of predicate argument structures from technical texts using Machine Learning: the system ASIUM. In Dieter Fensel Rudi Studer, editor, 11th European Workshop EKAW'99, pages 329-334, Springer-Verlag, Mai 1999.

  3. Faure, D. and Poibeau T. First experiments of using semantic knowledge learned by ASIUM for information extraction task using INTEX. In Staab, S. and Maedche, A. and Nedellec, C. and Wiemer-Hastings P., editors, Ontology Learning ECAI-2000 Workshop, pages 7-12, 22 Août 2000.

back to index