Brydon Eastman | Applied Math, University of Waterloo
Machine Learning Techniques in Mathematical Oncology
Cell signaling in early tumour development is crucial to the geometry of the invading solid tumour. This complex process is impacted by numerous factors in the cellular micro-environment including the presence of cancerous stem cells, methods of cell signaling, and nutrient fields. All these factors influence the proliferation and invasive potential of a cancerous tumour. We propose, and investigate, a selection of models that utilise partial differential equations, stochastic processes, and machine learning in hybrid techniques. The processes proposed seek to use machine learning in the context of artificial neural networks to identify cancer stem cells via image flow cytometery, to act as decision making units in cellular automata simulations of cancer stem cells, and to perform sensitivity and bifurcation analysis on bio-systems in mathematical oncology.