Special Session Introduction
Climate change is emerging as the single most significant challenge facing humanity today and is likely to influence and shape our societies into the 21st century and beyond. The heat-trapping nature of excessive amounts of Carbon Dioxide (CO2) and other gases generated as a result of human activity over the last 150 years, has caused a steady rise in average global temperatures. Scientist have determined that the long term effects of warming temperatures is causing shifts in weather patterns such as more frequent heatwaves, droughts and heavy precipitation while rising sea levels due to melting sea ice are increasing the risk of flooding in low laying coastal regions. The effects of climate change will be felt in every facet of society from those residing in large metropolitan areas to small rural communities and complicated by the growth of urbanisation, socio-economic and political factors.
In order to better understand, mitigate and reverse the effect of climate change there is a need to sense, capture and analyse heterogeneous data sources for monitoring its immediate and longer-term effects. In a recent paper on using Machine Learning (ML) for tackling the effects of climate changes (Rolnick et al 2019) researchers and industrial leaders in the field of artificial intelligence such as Google and DeepMind have outlined critical areas where the development and application of ML approaches can offer a range of possible solutions. More effective and innovate pervasive data-driven ML models and integrated systems need to be developed that could among other solutions:
- Predict and simulate extreme weather events and prescribe remedial measures such as the maintenance and deployment of assets (equipment) and municipal services.
- Develop effective monitoring of emissions and strategies for CO2 offset and removal.
- Manage renewable/hybrid energy production and power distribution.
- Pre-empt immediate or longer-term disasters and their effects such as flash flooding, forest fires, erosion and the effects of atmospheric, water, soil-based pollutants and toxicity levels.
- Monitor and optimise rural land use for agricultural production and ecological conservation.
- Optimise greener design and operation of physical infrastructures such as urban dwellings/facilities, land, air and sea based transportation and industrial production (to manage compounding environmental impacts such as urban heat islands).
- Monitor and simulate societal and economic impacts of climate changes at macro and micro scales.
- Model and predict the behaviour and effects of new approaches for climate engineering and solar radiation management.
To facilitate solutions for managing these and many other effects of climate change there is a need to leverage the power of biologically inspired ML methodologies for handling highly complex, uncertain and stochastic problems to create explainable insights and integrated systems. Computational intelligence (CI) approaches such as artificial neural networks can enable feature extraction, predictive modelling and generative solution creation using high dimensional multi-modal data sources. Fuzzy logic is an established methodology for handling imprecise and uncertain data (sensory, user-defined concepts, subjective opinions) providing an approach for approximate reasoning and modelling based on the use of linguistic quantifiers (fuzzy sets) and human interpretable fuzzy rules that can be used for inference and decision-making. Evolutionary algorithms are based on the process of natural selection, collective and cooperative intelligence for modelling stochastic systems and approaches which can be used for optimising complex real-world systems and processes. These core techniques together with new emerging approaches provide a basis on which their co-design and hybridisation in combination with relevant data sources can be used to provide effective modelling, simulation and control solutions for tackling the real-world challenges of climate change.
This IEEE WCCI special session aims to being together academic and industrial researchers to discuss how ML and CI techniques can be used to help solve challenging climate change problems as well as develop new approaches through the combination of potentially allied technologies such as embedded artificial intelligence, IoT, edge computing intelligence augmentation virtual and augmented reality.