My master’s thesis is a design framework for computational analysis Hollywood-style cinematic storytelling.
A use case for the development of a human readable data architecture for parsing film and a sketch for a probabilistic algorithm for teaching it to neural networks.
VIMEO PASSWORD: FRAMEWORK2016
catherine_Rehwinkel_Thesis =
[{a:['A computational, Bayesian model for film theory',
{b:{‘a use case': 'a conceptual design for a science-fiction universe', {'an android protagonist's: [' embodiment', 'algorithms', 'understanding of human perception of causality through film language and film narrative', {who:('experiences the world as if she is watching a film', 'expresses probabilistic/persuasive courses of action', {to:(human peers, family members)}, {via:('film language', 'narrative structure')})}}}},
{c:{design:{‘a hierarchical data architecture for machine learning to': 'parse film as an n-dimensional film-frame index’}}},
{i:{‘algorithms': {1: 'trains a neural network on director intention and audience reaction'}, {2: 'parses the features correlated with the persuasive success of a film’}}}}];