Improved N-Best Extraction with an Evaluation on Language Data

Abstract

We show that a previously proposed algorithm for the N-best trees problem can be made more efficient by changing how it arranges and explores the search space. Given an integer N and a weighted tree automaton (wta) M over the tropical semiring, the algorithm computes N trees of minimal weight with respect to M. Compared to the original algorithm, the modifications increase the laziness of the evaluation strategy, which makes the new algorithm asymptotically more efficient than its predecessor. The algorithm is implemented in the software Betty, and compared to the state-of-the-art algorithm for extracting the N best runs, implemented in the software toolkit Tiburon. The data sets used in the experiments are wtas resulting from real-world natural language processing tasks, as well as artificially created wtas with varying degrees of nondeterminism. We find that Betty outperforms Tiburon on all tested data sets with respect to running time, while Tiburon seems to be the more memory-efficient choice.

Generating Semantic Graph Corpora with Graph Expansion Grammar

“…Each output graph is saved as a single text file in gv format (one such file resulting from our running example is depicted in Figure 9), which makes their visualisation by Graphviz easy. We recommend using the Graphviz online tool 6 for quick and easy graph visualisation. An example of a visualisation of the graph in Figure 9 by Graphviz is shown in Figure 10.…”

“…Another example is Tiburon [13], a capable toolkit for processing weighted automata which includes an algorithm for extracting N parse trees with optimal weight from a weighted string grammar. Finally we have BETTY [6], which can extract both the N best derivation trees, but also the N best output trees with respect to a tree grammar; cf. Section 3.…”