Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Alberta

$25,000.00

Explainable Artificial Intelligence that Understands Human Language

FY 2023-24

Summary
In an increasingly data--driven world, large volumes of data are produced daily in many areas of society. It is now beyond the reach of human ability to extract and process information in a reliable and timely way, which makes it difficult to manually derive new knowledge and methods in support of data--driven decision making. Therefore, natural language processing (NLP) techniques are required that can automatically obtain knowledge from the daily volume of human--created data. As deep learning neural network methods gain in popularity as algorithms in NLP, the neural network's unexplainability becomes problematic, however. Current neural network models have a limitation that they cannot provide explanation on their decisions. Explanation on the prediction of a neural network model is critical for high stakes tasks (e.g., health and law) where people need to understand the artificial intelligence (AI) system's decisions. This research proposal aims to address the critical shortcoming of current, human-language related Explainable AI (XAI). The current XAI algorithms focus on the co-occurrence of words in a sentence, rather than on understanding the meaning of the input and providing an explanation based on their semantic understanding. My long-term vision is to develop algorithms for an explainable AI system that semantically understands human language. The impact will be that a growing number of Canadian professionals in high stakes domains would have the benefit of using explainable AI systems that can produce a prediction of an outcome, along with an explanation based on the semantic understanding of the data. To achieve this vision, my proposed research program has two long--term goals: My first long-term research goal, G1, is to create advanced algorithms and an integrated platform for explainable AI methods. For this goal, semantically structured representation of the input will be developed and tested. My second long--term research goal, G2, is to test automated explanation strategies for various scenarios and applications, and to identify the interpretability and explainability for end users as well as limitations in the real world. My research program will advance state--of--the--art explanations of black--box AI systems by creating highly reliable and robust systems. We will apply our systems to law and healthcare domains in which explanation of the prediction is critical. By using such adaptable systems, the use of AI will become more intuitive and reliable for Canadian legal and healthcare fields.
Program
Discovery Grants Program - Individual
Location
Alberta, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2022-03469
Name nom
Kim, Mi-Young
Department département
Augustana Campus
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2022
Award amount
25000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
9999
Research subject group en
Not available
Research subject group fr
Not available
Research subject en
Non disponible
Research subject fr
Non disponible
Application title
Explainable Artificial Intelligence that Understands Human Language
Keyword
Knowledge Acquisition; Natural Language Processing; Semantic Inference and Reasoning; Semantic Understanding of Language; Textual Entailment
Application summary
In an increasingly data--driven world, large volumes of data are produced daily in many areas of society. It is now beyond the reach of human ability to extract and process information in a reliable and timely way, which makes it difficult to manually derive new knowledge and methods in support of data--driven decision making. Therefore, natural language processing (NLP) techniques are required that can automatically obtain knowledge from the daily volume of human--created data. As deep learning neural network methods gain in popularity as algorithms in NLP, the neural network's unexplainability becomes problematic, however. Current neural network models have a limitation that they cannot provide explanation on their decisions. Explanation on the prediction of a neural network model is critical for high stakes tasks (e.g., health and law) where people need to understand the artificial intelligence (AI) system's decisions. This research proposal aims to address the critical shortcoming of current, human-language related Explainable AI (XAI). The current XAI algorithms focus on the co-occurrence of words in a sentence, rather than on understanding the meaning of the input and providing an explanation based on their semantic understanding. My long-term vision is to develop algorithms for an explainable AI system that semantically understands human language. The impact will be that a growing number of Canadian professionals in high stakes domains would have the benefit of using explainable AI systems that can produce a prediction of an outcome, along with an explanation based on the semantic understanding of the data. To achieve this vision, my proposed research program has two long--term goals: My first long-term research goal, G1, is to create advanced algorithms and an integrated platform for explainable AI methods. For this goal, semantically structured representation of the input will be developed and tested. My second long--term research goal, G2, is to test automated explanation strategies for various scenarios and applications, and to identify the interpretability and explainability for end users as well as limitations in the real world. My research program will advance state--of--the--art explanations of black--box AI systems by creating highly reliable and robust systems. We will apply our systems to law and healthcare domains in which explanation of the prediction is critical. By using such adaptable systems, the use of AI will become more intuitive and reliable for Canadian legal and healthcare fields.