FLSA: Extending Latent Semantic Analysis with Features for Dialogue Act Classification

FLSA: Extending Latent Semantic Analysis with features for dialogue act classification Riccardo Serafin CEFRIEL Via Fucini 2 20133 Milano, Italy Riccardo.Serafin@students.cefriel.it Barbara Di Eugenio Computer Science University of Illinois Chicago, IL 60607 USA bdieugen@cs.uic.edu Abstract We discuss Feature Latent Semantic Analysis (FLSA), an extension to Latent Semantic Analysis (LSA). LSA is a statistical method that is ordinar- ily trained on words only; FLSA adds to LSA the richness of the many other linguistic features that a corpus may be labeled with. We applied FLSA to dialogue act classification with excellent results. We report results on three corpora: CallHome Span- ish, MapTask, and our own corpus of tutoring dia- logues. 1 Introduction In this paper, we propose Feature Latent Semantic Analysis (FLSA) as an extension to Latent Seman- tic Analysis (LSA). LSA can be thought as repre- senting the meaning of a word as a kind of average of the meanings of all the passages in which it ap- pears, and the meaning of a passage as a kind of average of the meaning of all the words it contains (Landauer and Dumais, 1997). It builds a semantic space where words and passages are represented as vectors. LSA is based on Single Value Decompo- sition (SVD), a mathematical technique that causes the semantic space to be arranged so as to reflect the major associative patterns in the data. LSA has been successfully applied to many tasks, such as as- sessing the quality of student essays (Foltz et al., 1999) or interpreting the student’s input in an Intel- ligent Tutoring system (Wiemer-Hastings, 2001). A common criticism of LSA is that it uses only words and ignores anything else, e.g. syntactic in- formation: to LSA, man bites dog is identical to dog bites man. We suggest that an LSA semantic space can be built from the co-occurrence of arbitrary tex- tual features, not just words. We are calling LSA augmented with features FLSA, for Feature LSA. Relevant prior work on LSA only includes Struc- tured Latent Semantic Analysis (Wiemer-Hastings, 2001), and the predication algorithm of (Kintsch, 2001). We will show that for our task, dialogue act classification, syntactic features do not help, but most dialogue related features do. Surprisingly, one dialogue related feature that does not help is the di- alogue act history. We applied LSA / FLSA to dialogue act classi- fication. Dialogue systems need to perform dia- logue act classification, in order to understand the role the user’s utterance plays in the dialogue (e.g., a question for information or a request to perform an action). In recent years, a variety of empiri- cal techniques have been used to train the dialogue act classifier (Samuel et al., 1998; Stolcke et al., 2000). A second contribution of our work is to show that FLSA is successful at dialogue act classi- fication, reaching comparable or better results than other published methods. With respect to a baseline of choosing the most frequent dialogue act (DA), LSA reduces error rates between 33% and 52%, and FLSA …

Content truncated to 500 words