Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

McMaster University

$36,000.00

Evolving Creative Algorithms

FY 2024-25

Summary
Intelligent behaviour emerges in natural systems as a result of adaptation occurring over two vastly different timescales: evolution and lifetime learning. Both modes of adaptation are wildly creative, and their interaction is critical to developing robust intelligence. However, evolution and learning are often studied independently in nature-inspired AI. Furthermore, recent advances driven by deep neural networks have primarily been limited to specialized behaviours that excel at a single problem in a stationary environment. The models are computationally demanding and their training requires specialized hardware and significant energy resources. By contrast, natural nervous systems are highly efficient and emerged in part because they evolved in non-stationary, multi-task environments where survival required managing multiple objectives and modes of behaviour over time (e.g. predator and prey). The long-term objective of this research is to synergize evolutionary model building and computational neuroscience to discover algorithms that think, learn, and create like people as they solve non-stationary and multi-task problems. Creativity, a learner's capacity to generate new things of value, is key to trial-and-error problem solving. It will be a fundamental property of Artificial General Intelligence in the future. The underlying approach of this research is the evolution of computer programs using natural selection, i.e. genetic programming. While genetic programming can theoretically evolve any algorithm from scratch using basic mathematical operations as building blocks, this flexibility presents a massive search space that makes discovering complex algorithms difficult. We will enhance the evolutionary search for intelligent behaviour by guiding genetic programming using knowledge about how brains perform perception, prediction, and action in creative ways. This opens short-term questions about how digital evolution can continuously discover new programs that incorporate neurobiological properties such as spatial sensorimotor connectivity, probabilistic behaviour with continual active learning, and open-ended hierarchical development. Multi-task prediction and control have a breadth of real-world applications that are critical to industry and society, from autonomous robotics to predicting an EEG trace to anticipate changes in patient health. A common property of most real-world environments is that things change. Creativity will allow intelligent machines to adapt to changing environments in an open-ended way: building solutions to new problems and building new ways of problem solving. This work will harness the tendency of Darwinian evolution to discover organisims that are creative and energy-efficient. Efficient AI is a vital response to the climate crisis, improves accessibility of AI research and STEM education, and decreases the cost of AI applications in society.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2023-05663
Name nom
Kelly, Stephen
Department département
Computing and Software
Organization id
27
Institution établissement
McMaster University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
36000
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
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
Research subject code
2805
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Learning and inference theories
Research subject fr
Théories de l'apprentissage et l'inférence
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Evolving Creative Algorithms
Keywords
genetic programming; evolutionary computation; computational neuroscience
Application summary
Intelligent behaviour emerges in natural systems as a result of adaptation occurring over two vastly different timescales: evolution and lifetime learning. Both modes of adaptation are wildly creative, and their interaction is critical to developing robust intelligence. However, evolution and learning are often studied independently in nature-inspired AI. Furthermore, recent advances driven by deep neural networks have primarily been limited to specialized behaviours that excel at a single problem in a stationary environment. The models are computationally demanding and their training requires specialized hardware and significant energy resources. By contrast, natural nervous systems are highly efficient and emerged in part because they evolved in non-stationary, multi-task environments where survival required managing multiple objectives and modes of behaviour over time (e.g. predator and prey). The long-term objective of this research is to synergize evolutionary model building and computational neuroscience to discover algorithms that think, learn, and create like people as they solve non-stationary and multi-task problems. Creativity, a learner's capacity to generate new things of value, is key to trial-and-error problem solving. It will be a fundamental property of Artificial General Intelligence in the future. The underlying approach of this research is the evolution of computer programs using natural selection, i.e. genetic programming. While genetic programming can theoretically evolve any algorithm from scratch using basic mathematical operations as building blocks, this flexibility presents a massive search space that makes discovering complex algorithms difficult. We will enhance the evolutionary search for intelligent behaviour by guiding genetic programming using knowledge about how brains perform perception, prediction, and action in creative ways. This opens short-term questions about how digital evolution can continuously discover new programs that incorporate neurobiological properties such as spatial sensorimotor connectivity, probabilistic behaviour with continual active learning, and open-ended hierarchical development. Multi-task prediction and control have a breadth of real-world applications that are critical to industry and society, from autonomous robotics to predicting an EEG trace to anticipate changes in patient health. A common property of most real-world environments is that things change. Creativity will allow intelligent machines to adapt to changing environments in an open-ended way: building solutions to new problems and building new ways of problem solving. This work will harness the tendency of Darwinian evolution to discover organisims that are creative and energy-efficient. Efficient AI is a vital response to the climate crisis, improves accessibility of AI research and STEM education, and decreases the cost of AI applications in society.