The Topology of Semantic Knowledge

Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 669–680, Seattle, Washington, USA, 18-21 October 2013. c⃝2013 Association for Computational Linguistics The Topology of Semantic Knowledge Jimmy Dubuisson Jean-Pierre Eckmann D´epartement de Physique Th´eorique and Section de Math´ematiques Universit´e de Gen`eve Jimmy.Dubuisson@unige.ch Christian Scheible Institut f¨ur Maschinelle Sprachverarbeitung University of Stuttgart scheibcn@ims.uni-stuttgart.de Hinrich Sch¨utze Center for Information and Language Processing University of Munich Abstract Studies of the graph of dictionary definitions (DD) (Picard et al., 2009; Levary et al., 2012) have revealed strong semantic coherence of local topological structures. The techniques used in these papers are simple and the main results are found by understanding the struc- ture of cycles in the directed graph (where words point to definitions). Based on our ear- lier work (Levary et al., 2012), we study a dif- ferent class of word definitions, namely those of the Free Association (FA) dataset (Nelson et al., 2004). These are responses by subjects to a cue word, which are then summarized by a directed, free association graph. We find that the structure of this network is quite different from both the Wordnet and the dictionary networks. This difference can be explained by the very nature of free associa- tion as compared to the more “logical” con- struction of dictionaries. It thus sheds some (quantitative) light on the psychology of free association. In NLP, semantic groups or clusters are inter- esting for various applications such as word sense disambiguation. The FA graph is tighter than the DD graph, because of the large num- ber of triangles. This also makes drift of meaning quite measurable so that FA graphs provide a quantitative measure of the seman- tic coherence of small groups of words. 1 Introduction The computer study of semantic networks has been around since the advent of computers (Brunet, 1974) and has been used to study semantic relations be- tween concepts and for analyzing semantic data. Traditionally, a popular lexical database of English is Wordnet (Miller, 1995; Miller and Fellbaum, 1998), which organizes the semantic network in terms of graph theory. In contrast to manual ap- proaches, the automatic analysis of semantically in- teresting graph structures of language has received increasing attention. For example, it has become clear more recently that cycles and triangles play an important role in semantic networks, see e.g., (Dorow et al., 2005). These results suggest that the underlying semantic structure of language may be discovered through graph-theoretical methods. This is in line with similar findings in much wider realms than NLP (Eckmann and Moses, 2002). In this paper, we compare two different types of association networks. The first network is con- structed from an English dictionary (DD), the sec- ond from a free association (FA) database (Nelson et al., 2004). We represent both datasets through directed graphs. For DD, the nodes are words and the directed edges point from a word to its defini- tion(s). For FA, the nodes are again words, and each cue word has a directed edge …

Content truncated to 500 words