Split-Based Algorithm for Weighted Context-Free Grammar Induction

Abstract

The split-based method in a weighted context-free grammar (WCFG) induction was formalised and verified on a comprehensive set of context-free languages. WCFG is learned using a novel grammatical inference method. The proposed method learns WCFG from both positive and negative samples, whereas the weights of rules are estimated using a novel Inside–Outside Contrastive Estimation algorithm. The results showed that our approach outperforms in terms of F1 scores of other state-of-the-art methods.

Introducing a Pure Weight-Based Concept for Automatic Test Case Generators

“…A weighted context-free grammar (or WCFG, as described in [7]) can be defined as a 5-tuple 𝐺 = (𝑇 𝐺 , 𝑁 𝐺 , 𝑃 𝐺 , 𝑆 𝐺 , 𝑊 𝐺 ), where 𝑊 𝐺 has a finite relation 1:1 with 𝑃 𝐺 (wi identifies the weight set for production pi).…”

“…Moreover, considering all 𝑝 𝑖 expanding the same symbol, within our weight-based approach the sum of corresponding wi must always be 1, making this equivalent to the definition of probabilistic context-free grammar (PCFG [7], [8], equally expressive as WCFG); using the function ℛ introduced in Section 2, such constraint can be expressed as ∀𝐴 ∈ 𝑁 𝐺 : ∑ 𝑤 𝑖 = 1…”