PRESS RELEASE | For immediate release
CRIM takes part in a vast technological project on Canadian Aboriginal languages
Montréal, December 5, 2018 – CRIM is proud to announce the start of a long-term collaboration with the National Research Council of Canada (NRC) on a project to encourage the revitalization and preservation of Aboriginal languages through text and speech-based technologies. CRIM will contribute its expertise in adapting its speech recognition technologies to Aboriginal languages.
The Technologies for Canadian Aboriginal Languages Project is a CRNC initiative funded by the federal government in Budget 2017. CRIM will also be working with researchers from Carleton University and the University of Alberta who have been involved with Aboriginal communities for many years. The teams will work in partnership with Aboriginal community organizations and Aboriginal communities across Canada.
Our team is delighted to be collaborating with CRIM, which is world-renowned for its text- and speech-based technologies, on this important cultural and linguistic challenge.
– Roland Kuhn, NRC
CRIM’s role
CRIM will carry out two main projects that will form the basis for the development of a dozen speech recognition systems adapted to the target languages.
- The first project will lead to the development of tools for segmenting audio recordings, making it possible to distinguish speech from music or noise, identify spoken languages and separate different speakers. These tools will facilitate the annotation and transcription of content to speed up the documentation of existing corpora for each language.
- The second project focuses on the design of an indexing tool to identify and organize existing audio content for each of the target languages. This will make it easy to navigate through the many recordings available, for example to find video sequences where a specific topic is being discussed, or to discover how certain expressions are used.
A crucial project, both in terms of its social impact and scientific progress
Researcher Roland Kuhn, leader of the Canadian Aboriginal Language Technologies Project at NRC, describes why this joint research with CRIM is important for Canada’s aboriginal communities: “Thousands of hours of aboriginal language recordings exist, but unfortunately they are rarely annotated or indexed due to the lack of technology to do so: this unannotated corpus continues to grow. This situation is frustrating for members of the communities concerned, who would like to be able to search using keywords to find recordings relevant to their current needs. A person shouldn’t have to listen to 10,000 hours of audio to hope to find a recording of a traditional ceremony in their community, for example. Our team is very pleased to be collaborating with CRIM, which is world-renowned for its text and speech-based technologies, on this important cultural and linguistic challenge.”
The structure of native languages is very different from that of English or French, making many of today’s speech recognition methods inadequate. This poses a major challenge and calls for the development of new approaches that will certainly lead to innovations in the field.
Both parties aim to provide quality technological tools for language teaching, enhancement and preservation for aboriginal communities. CRIM experts also hope that the methods developed for the languages targeted at the start of the project (Inuktitut and Cree) can be applied to many other of the 70 Aboriginal languages spoken in Canada.
About CRIM
CRIM is an applied research and expertise centre in information technology, dedicated to making organizations more effective and competitive through the development of innovative technology and the transfer of leading edge know-how, while contributing to scientific advancement.
It enables organizations, mainly SMEs, to demystify and gain access to cutting-edge technologies such as artificial intelligence, in order to effectively resolve the technological issues they face. Its researchers and IT professionals are developing a wide range of applications in diverse sectors, working in fields of expertise such as machine learning, computer vision, speech recognition, automatic natural language processing, data science and operations research.
As a non-profit organization, CRIM’s neutrality and the strength of its network make it an indispensable resource. Its actions are in line with the policies and strategies steered by the Ministère de l’Économie et de l’Innovation, its main financial partner.
– 30 –
Source: CRIM
