Aligning the IndoWordNet with the Princeton WordNet

The IndoWordNet is an Indian language lexical resource. The project started with HindiWordNet, which was manually built from various resources with the preference for culture-specific synsets. Other languages were added later. The development approach used in IndoWordNet is very similar to that used in Princeton WordNet (PWN). PWN is a semantic network where English synsets are nodes, and semantic relations are edges connecting them. Due to the popularity of PWN, IndoWordNet also connected Hindi and English languages through direct (synonymy) and hypernymy linkages between their synsets. Due to the diversity of the languages, these linkages generate three types of mappings between IndoWordNet and PWN which generate the misalignment. This paper proposes to align the IndoWordNet with PWN using a largescale lexical-semantic resource called Universal Knowledge Core (UKC), which forms asemantic network where nodes are language-independent concepts. In the UKC semantic relations connect concepts and not synsets.