DCS240

Biological intelligence is characterized by selecting, processing, and storing information while flexibly adapting to changing conditions. How might biology inspire “smart” algorithms? This course explores the fundamental principles of artificial neural networks (ANNs). Students begin with modeling learning in a single computational unit (McCulloch-Pitts neuron), and then examine how many simple units can collectively give rise to complex behaviors. They examine both supervised networks that learn a predetermined input-output relationship, and unsupervised networks that learn “suspicious coincidences” from the input data. They implement neural networks with Python (previous experience is helpful but not necessary). Prerequisite(s), which may be taken concurrently: NS/PY 160. Recommended background: Experience with Python programming (such as from DCS 109) would be helpful, but is not strictly required.