[Paper Review] Parsing with Typed Feature Structures

Shuly Wintner, Nissim Francez|arXiv (Cornell University)|Jan 31, 1996

Natural Language Processing Techniques3 citations

TL;DR

This paper presents a formal framework for parsing using Typed Feature Structures (TFS) within a restricted version of the ALE formalism, introducing Abstract Typed Feature Structures (AFSs) and Multi-Rooted Feature Structures (MRSs) to model grammatical rules and phrasal signs. It defines unification over AFSs and implements a correct bottom-up chart parser, establishing a foundation for an abstract machine based on operational semantics.

ABSTRACT

In this paper we provide for parsing with respect to grammars expressed in a general TFS-based formalism, a restriction of ALE. Our motivation being the design of an abstract (WAM-like) machine for the formalism, we consider parsing as a computational process and use it as an operational semantics to guide the design of the control structures for the abstract machine. We emphasize the notion of abstract typed feature structures (AFSs) that encode the essential information of TFSs and define unification over AFSs rather than over TFSs. We then introduce an explicit construct of multi-rooted feature structures (MRSs) that naturally extend TFSs and use them to represent phrasal signs as well as grammar rules. We also employ abstractions of MRSs and give the mathematical foundations needed for manipulating them. We then present a simple bottom-up chart parser as a model for computation: grammars written in the TFS-based formalism are executed by the parser. Finally, we show that the parser is correct.

Motivation & Objective

• To develop a formal computational model for parsing grammars expressed in a TFS-based formalism.
• To design an abstract machine inspired by the Warren Abstract Machine (WAM) for efficient execution of TFS-based grammars.
• To define unification over abstract typed feature structures (AFSs) rather than concrete TFSs to simplify computation.
• To introduce multi-rooted feature structures (MRSs) as a natural extension of TFSs for representing phrasal signs and grammar rules.
• To provide mathematical foundations for manipulating AFSs and MRSs to support parsing operations.

Proposed method

• The paper introduces Abstract Typed Feature Structures (AFSs) as a computational abstraction of TFSs, focusing on essential information for parsing.
• It defines unification over AFSs to support efficient and correct inference in the parsing process.
• Multi-Rooted Feature Structures (MRSs) are introduced as an extension of TFSs to represent multiple syntactic heads in a single structure.
• The formalism uses MRSs to encode both phrasal signs and grammar rules in a unified way.
• A bottom-up chart parser is designed as a computational model, executing TFS-based grammars via unification and chart-based search.
• The correctness of the parser is formally proven using operational semantics derived from the parsing process.

Experimental results

Research questions

• RQ1 How can typed feature structures be abstracted to support efficient and correct parsing in a formal computational model?
• RQ2 What is the role of multi-rooted feature structures (MRSs) in representing phrasal signs and grammar rules?
• RQ3 How can unification be defined over abstract typed feature structures (AFSs) to support parsing without full TFS instantiation?
• RQ4 What mathematical foundations are needed to manipulate AFSs and MRSs in a way that supports parsing?
• RQ5 How can a bottom-up chart parser be formally specified and proven correct for TFS-based grammars?

Key findings

• The paper successfully defines a formal unification mechanism over Abstract Typed Feature Structures (AFSs), enabling efficient parsing without full TFS manipulation.
• Multi-Rooted Feature Structures (MRSs) are shown to naturally represent both phrasal signs and grammar rules within a single formalism.
• The proposed bottom-up chart parser is formally proven correct for TFS-based grammars, establishing its reliability as a computational model.
• The framework provides a solid mathematical foundation for manipulating AFSs and MRSs, supporting the design of an abstract machine for TFS-based computation.
• The work establishes a clear operational semantics for TFS-based grammars, guiding the design of a WAM-like abstract machine.