Return to the main site:
Artificial Neural Networks (ANNs) are a class of machine learning methods based on features of biological brains, such as reinforcement and massive parallelism. These computational models are highly general, with applications to many classes of scientific and engineering problems, including pattern recognition, classification, and adaptive control. The objective of the course is for students to obtain a working knowledge of this forefront technology. This class provides a conceptual introduction to the study of ANNs within the wider field of Artificial Intelligence. We cover theoretical foundations, capabilities, and limitations, as well as methods from supervised, unsupervised, and reinforcement learning. A theme of the class is its connection to biology, where the comparison of artificial and natural systems informs and inspires both domains from the level of the neuron to the definition of intelligence. A second theme is the relationship between ANNs and other aspects of systems science, such as adaptation, complexity, emergence, and control theory.