Coarse-to-fine <i>n</i>-best parsing and MaxEnt discriminative reranking

TLDR

Discriminative reranking constructs high-performance statistical parsers but requires a source of candidate parses for each sentence. The paper proposes a novel method to generate 50-best parse sets using a coarse-to-fine generative parser. The method first produces 50-best parse sets with a coarse-to-fine generative parser, then reranks them with a MaxEnt reranker to select the best parse. The approach yields 50-best lists of substantially higher quality, achieving a 91.0 % f‑score on sentences up to length 100.

Abstract

Discriminative reranking is one method for constructing high-performance statistical parsers (Collins, 2000). A discriminative reranker requires a source of candidate parses for each sentence. This paper describes a simple yet novel method for constructing sets of 50-best parses based on a coarse-to-fine generative parser (Charniak, 2000). This method generates 50-best lists that are of substantially higher quality than previously obtainable. We used these parses as the input to a MaxEnt reranker (Johnson et al., 1999; Riezler et al., 2002) that selects the best parse from the set of parses for each sentence, obtaining an f-score of 91.0% on sentences of length 100 or less.