Semantic Parsing Using Content and Context: A Case Study from Requirements Elicitation

Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1296–1307, October 25-29, 2014, Doha, Qatar. c⃝2014 Association for Computational Linguistics Semantic Parsing Using Content and Context: A Case Study from Requirements Elicitation Reut Tsarfaty Weizmann Institute Rehovot, Israel Ilia Pogrebezky Interdisciplinary Center Herzliya, Israel Guy Weiss Weizmann Institute Rehovot, Israel Yaarit Natan Weizmann Institute Rehovot, Israel Smadar Szekely Weizmann Institute Rehovot, Israel David Harel Weizmann Institute Rehovot, Israel Abstract We present a model for the automatic se- mantic analysis of requirements elicitation documents. Our target semantic repre- sentation employs live sequence charts, a multi-modal visual language for scenario- based programming, which can be directly translated into executable code. The ar- chitecture we propose integrates sentence- level and discourse-level processing in a generative probabilistic framework for the analysis and disambiguation of individual sentences in context. We show empiri- cally that the discourse-based model con- sistently outperforms the sentence-based model when constructing a system that re- flects all the static (entities, properties) and dynamic (behavioral scenarios) require- ments in the document. 1 Introduction Requirements elicitation is a process whereby a system analyst gathers information from a stake- holder about a desired system (software or hard- ware) to be implemented. The knowledge col- lected by the analyst may be static, referring to the conceptual model (the entities, their properties, the possible values) or dynamic, referring to the behavior that the system should follow (who does what to whom, when, how, etc). A stakeholder in- terested in the system typically has a specific static and dynamic domain in mind, but he or she cannot necessarily prescribe any formal models or code artifacts. The term requirements elicitation we use here refers to a piece of discourse in natural lan- guage, by means of which a stakeholder commu- nicates their desiderata to the system analyst. The role of a system analyst is to understand the different requirements and transform them into formal constructs, formal diagrams or executable code. Moreover, the analyst needs to consolidate the different pieces of information to uncover a single shared domain. Studies in software engi- neering aim to develop intuitive symbolic systems with which human agents can encode require- ments that would then be unambiguously trans- lated into a formal model (Fuchs and Schwitter, 1995; Bryant and Lee, 2002). More recently, Gordon and Harel (2009) de- fined a natural fragment of English that can be used for specifying requirements which can be effectively translated into live sequence charts (LSC) (Damm and Harel, 2001; Harel and Marelly, 2003), a formal language for specifying the dynamic behavior of reactive systems. How- ever, the grammar that underlies this language fragment is highly ambiguous, and all disam- biguation has to be conducted manually by a hu- man agent. Indeed, it is commonly accepted that the more natural a controlled language fragment is, the harder it is to develop an unambiguous translation mechanism (Kuhn, 2014). In this paper we accept the ambiguity of re- quirements descriptions as a premise, and aim to …

Content truncated to 500 words