Graph transformation for incremental natural language analysis

Abstract

Millstream systems have been proposed as a non-hierarchical method for modelling natural language. Millstream configurations represent and connect multiple structural aspects of sentences. We present a method by which the Millstream configurations corresponding to a sentence are constructed. The construction is incremental, that is, it proceeds as the sentence is being read and is complete when the end of the sentence is reached. It is based on graph transformations and a lexicon which associates words with graph transformation rules that implement the incremental construction process. Our main result states that, for an effectively nonterminal-bounded reader R and a Millstream system MS based on monadic second-order logic, the correctness of R with respect to MS can be checked: it is decidable whether all graphs generated by R belong to the language of configurations specified by MS .

Towards proactive robot behavior based on incremental language analysis

“…output from a speech-to-text module in the robot. For the incremental analysis we use a Millstream reader [1], which transforms a sentence into a so called Millstream configuration, comprising representations (also referred to as modules) for syntax, semantics, and expected robot action, and the dependencies between these three components. In this paper the syntactic structure describes the constituency relation of a sentence.…”

“…Throughout the analysis, partial results, available in the current configurations, can be partially acted upon by the robot as exemplified in the introduction. As mentioned in [1], one important aspect for future work is the degree of nondeterminism, that is, how to handle several partial results. Future work also includes grounding of action parameters to physical objects that the robot perceives.…”

“…[6]). Incremental processing of speech has advantages both for the language analysis as such [1] and for the inference of robot actions. Partial results provide a possibility for the robot to react proactively, for instance in terms of sensing, perception, and physical motion.…”

Proceedings of the SIGFSM Workshop on Statistical NLP and Weighted Automata

“…The flexibility of aligned HR bimorphisms goes well beyond hybrid grammars as they generalize the synchronous HR grammars of (Jones et al, 2012), making it possible to synchronously generate two graphs connected by explicit alignment structures. Thus, they can for instance model alignments involving directed acyclic graphs like Abstract Meaning Representations (Banarescu et al, 2013) or Millstream systems (Bensch et al, 2014).…”