WC130 MayJune2023 - Magazine - Page 11
as well as ambient air temperature.
Where conventional wisdom uses
Canada is in a unique position to be a global leader in
the extrapolation of recent trends to
the use of artificial intelligence and machine learning
predict future outcomes, AI and ML
models base future outcomes on the
models within the water industry.
relationship between multiple parameters and actual observed results.
Many regulatory agencies require
four seasons of monitoring at a minimum, sometimes over multiple years, in advance of
approval for work. Rather than maintaining the status quo, we should embrace the use of technological
advances in AI and ML models within the approvals
process to demonstrate future outcomes and expedite
the approval process for low-risk activities. Using AI
and ML models in the approvals process could reduce
the burden on an applicant as well as free-up resources within the regulatory agencies, allowing for more
time to focus on sensitive or high-risk activities.
Canada is in a unique position to be a global
leader in the use of artificial intelligence and machine
learning models within the water industry. We should
be embracing the technology and our abundance of
publicly available data as a means to improve all facets of the operation of water systems, the regulatory
approval process, and watershed management.
Putting the ML in AI
Getty Images
The term artificial intelligence is a broad one that
includes the many processes that utilize computers to
mimic the cognitive abilities and functions of humans,
ultimately combining computer science and datasets to
enable problem-solving. Machine learning is a subset
of AI technology that focuses specifically on the ability
of the machine or computer to receive data and predict
future outcomes, while continuously improving the
model accuracy as new data is received.
The ML algorithm (model) learns from the various
combinations of parameters in historical data and uses
those relationships to estimate future results. It is an
intelligent tool that learns from past data with the end
goal of making better predictions of future outcomes
with a higher degree of certainty. The ML model divides
historical data into training and testing sets and then
evaluates the data sets by comparing the outcomes
of the model with the actual data in the test data set.
This process is iterative and repeated until there is a
significant level of confidence in the model’s predictive
capability. At that point, the model can be used for
future predictions.
WAT E R C A N A D A . N E T
WATER C AN ADA • M AY/JUNE 2023
11
