Peter Ljunglöf

GoDiS: an accommodating dialogue system

This paper accompanies a demo of the GoDiS system. Work on this system was reported at IJCAI-99 (Bohlin et al., 1999). GoDiS is a prototype dialogue system for information-seeking dialogue, capable of accommodating questions and tasks to enable the user to present information in any desired order, without explicitly naming the dialogue task. GoDiS is implemented using the TRINDIKIT software package, which enables implementation of these behaviours in a compact and natural way.

Typed Logical Variables in Haskell

We describe how to embed a simple typed functional logic programming language in Haskell. The embedding is a natural extension of the Prolog embedding by Seres and Spivey. To get full static typing we need to use the Haskell extensions of quantified types and the ST-monad.

Fast Statistical Parsing with Parallel Multiple Context-Free Grammars

We present an algorithm for incremental statistical parsing with Parallel Multiple Context-Free Grammars (PMCFG). This is an extension of the algorithm by Angelov (2009) to which we added statistical ranking. We show that the new algorithm is several times faster than other statistical PMCFG parsing algorithms on real-sized grammars. At the same time the algorithm is more general since it supports non-binarized and non-linear grammars. We also show that if we make the search heuristics non-admissible, the parsing speed improves even further, at the risk of returning sub-optimal solutions.

A Grammar Formalism for Specifying ISU-Based Dialogue Systems

We describe how to give a full specification of an ISU-based dialogue system as a grammar. For this we use Grammatical Framework (GF), which separates grammars into abstract and concrete syntax. All components necessary for a working GoDiS dialogue system are specified in the abstract syntax, while the linguistic details are defined in the concrete syntax. Since GF is a multilingual grammar formalism, it is straightforward to extend the dialogue system to several languages. Furthermore, the GF Resource Grammar Library can be used to write a single concrete instance covering 13 different languages.

A polynomial time extension of parallel multiple context-free grammar

It is already known that parallel multiple context-free grammar (PMCFG) [1] is an instance of the equivalent formalisms simple literal movement grammar (sLMG) [2, 3] and range concatenation grammar (RCG) [4, 5]. In this paper we show that by adding the single operation of intersection, borrowed from conjunctive grammar [6], PMCFG becomes equivalent to sLMG and RCG. As a corollary we get that PMCFG with intersection describe exactly the class of languages recognizable in polynomial time.

ShrdLite: Semantic Parsing Using a Handmade Grammar

This paper describes my approach for parsing robot commands, which was task 6 at SemEval 2014. My solution is to manually create a compact unification grammar. The grammar is highly ambiguous, and relies heavily on filtering the parse results by checking their consistency with the current world. The grammar is small, consisting of not more than 25 grammatical and 60 lexical rules. The parser uses simple error correction together with a straightforward iterative deepening search. Nevertheless, with these very basic algorithms, the system still managed to get 86.1% correctness on the evaluation data. Even more interesting is that by making the parser slightly more robust, the accuracy of the system rises to 93.5%, and by adding one single word to the lexicon, the accuracy is boosted to 98.0%.

Converting Grammatical Framework to Regulus

We present an algorithm for converting Grammatical Framework grammars (Ranta, 2004) into the Regulus unification-based framework (Rayner et al., 2006). The main purpose is to take advantage of the Regulusto-Nuance compiler for generating optimized speech recognition grammars. But there is also a theoretical interest in knowing how similar the two grammar formalisms are. Since Grammatical Framework is more expressive than Regulus, the resulting Regulus grammars can be overgenerating. We therefore describe a subclass of Grammatical Framework for which the algorithm results in an equivalent Regulus grammar.