National Research Council Canada (NRC) is seeking a postdoctoral candidate with an interest in ethical aspects of Natural Language Processing (NLP). The goal of the project is to develop tools and techniques to address two critical issues in ethical NLP applications design and use: fairness and explainability. AI applications often inadvertently perpetuate and accentuate unfair biases that can originate from multiple sources, such as training data, labeling process, data sampling, etc. Biased outputs can negatively affect certain demographic groups of users and even lead to discrimination. The ability of an AI system to provide understandable explanations for its decisions is a crucial factor in real-life applications both for developers to better understand the system's behavior and for users to gain trust in the system. The techniques will be developed with the focus on one application area, abusive language detection in social media, while ensuring their applicability and/or transferability to other natural language understanding tasks and domains.
The successful applicant will join the Text Analytics group at NRC Digital Technologies Research Centre, which has a world-renowned reputation in Natural Language Processing and Computational Linguistics, including sentiment and emotion analysis, information extraction, medical informatics and health-related applications. The applicant will work in close collaboration with the project leads, Dr. Svetlana Kiritchenko and Dr. Isar Nejadgholi, but will also have access to diverse expertise of other NRC researchers.
Position duration: 2 years
Start time: flexible
Location of position: NRC Montreal Road Campus, Ottawa, Ontario, Canada
How to apply: please send your CV and an expression of interest to svetlana <dot> kiritchenko <at> nrc-cnrc <dot> gc <dot> ca before July 1, 2020 for full consideration. After the deadline, the applications will be accepted until the position is filled.
The National Research Council Canada (NRC) is Canada's largest federal research and development organization. It brings together the brightest minds to deliver tangible impacts on the lives of Canadians and people around the world. The NRC's text analytics team conceives, builds, and evaluates tools that help extract information from textual data: explicit, implicit and inferred information, composite information (summaries), subtext information (sentiment, sarcasm, metaphor) and meta-information. The team collaborates with government, industry, and academia to drive text analytics technologies for a range of problem-driven and data-driven applications. The team has been working in areas such as situational awareness, clinical research informatics, sentiment and emotion analysis, and intelligence. Its innovative capacity and technical competency in machine learning methods and (big-) data-oriented methods have resulted in world-leading research, collecting multiple first-place rankings in text analysis research challenges such as i2b2 NLP for Clinical Data and the Sentiment Analysis track of SemEval.
At NRC, we know diversity enables excellence in research and innovation. We are committed to a diverse and representative workforce, an open and inclusive work environment, and contributing to a more inclusive Canadian innovation system.
Due to its rich ethnic, linguistic and religious diversity, and its strong talent base, Ottawa has emerged as an exciting cosmopolitan centre. A study by Canadian Business ranked Ottawa as Canada's best place to live based on incomes, car ownership, unemployment rate, housing, weather and lifestyle. Cultural factors were a key element of Ottawa's high ranking. Canadian Business magazine ranked Ottawa 14th in the world and 3rd in North America in a 2009 survey of the best places to live. In 2010, Ottawa was among the top 10 percent of cities globally in a 'quality of living survey' that included the economy, political system, education and schools, health, housing and the environment. Ottawa was rated as the best overall city in Canada based on a measure of technology, talent and tolerance.