Global Open Resources and Information for Language and Linguistic Analysis (GORILLA)

Global Open Resources and Information for Language and Linguistic Analysis (GORILLA) Damir Cavar, Małgorzata E. ´Cavar, Lwin Moe Indiana University Bloomington, IN, USA {dcavar,mcavar,lwinmoe}@indiana.edu Abstract The infrastructure Global Open Resources and Information for Language and Linguistic Analysis (GORILLA) was created as a resource that provides a bridge between disciplines such as documentary, theoretical, and corpus linguistics, speech and language technologies, and digital language archiving services. GORILLA is designed as an interface between digital language archive services and language data producers. It addresses various problems of common digital language archive infrastructures. At the same time it serves the speech and language technology communities by providing a platform to create and share speech and language data from low-resourced and endangered languages. It hosts an initial collection of language models for speech and natural language processing (NLP), and technologies or software tools for corpus creation and annotation. GORILLA is designed to address the Transcription Bottleneck in language documentation, and, at the same time to provide solutions to the general Language Resource Bottleneck in speech and language technologies. It does so by facilitating the cooperation between documentary and theoretical linguistics, and speech and language technologies research and development, in particular for low-resourced and endangered languages. Keywords: Speech and Language Corpora, NLP, Low-resourced Languages, Transcription Bottleneck 1. Introduction The Global Open Resources and Information for Lan- guage and Linguistic Analysis (GORILLA) project1 builds a cyber-infrastructure and platform that addresses numer- ous issues related to digital language resources, language documentation, and speech and language technologies. A particular focus of the project is on low-resourced or en- dangered languages. We assume that currently the majority of all languages in the world belongs to this group, since major resources in form of corpora, treebanks, transcribed and annotated speech data, or just dictionaries and formal grammars are missing.2 The goal of GORILLA is to bridge the gap between large amounts of available language documentation data in var- ious archives and repositories, and the lack of speech and language data for Natural Language Processing (NLP) or Human Language Technologies (HLT) research and devel- opment needs. We expect that documentation, transcription and annotation of available language resources can be im- proved, if NLP and HLT communities cooperate with docu- mentary linguists and language archives. At the same time, making language resources from low-resourced and endan- gered languages accessible to the NLP and HLT community will significantly improve common speech and language algorithms and technologies, and indirectly the entire lan- guage documentation and research discipline, as it already did for the most common 1% of high-resourced languages. 1See http://gorilla.linguistlist.org. 2Depending on the question and need, as Christopher Cieri (LDC) pointed out (p.c.), even high resources languages like En- glish could be considered low-resourced with respect to some re- source types that are missing currently or in the process of emerg- ing new technologies and approaches. The classification of low- resourced is a relative and dynamic one that needs to be put in a context. 2. Language Documentation Resources In their attempt to …

Content truncated to 500 words