How machines sense, reason, and learn — taught from the ground up.
CS 1010
LECTURE SERIES
16 WEEKS
Definition
Building systems that perform tasks we'd call intelligent — seeing, reading, deciding, predicting.
The shift that matters: modern AI isn't hand-coded rule by rule. It learns patterns from data, the way a student learns from worked examples.
Module 01
Before a system can recognize an image or translate a sentence, it has to learn from data. This module defines what "learning" actually means for a machine.
Module 01
Learns from labeled examples — inputs paired with correct answers — then predicts answers for new inputs.
No labels. The system explores raw data and finds its own structure — clusters, patterns, outliers.
Learns by doing — taking actions, earning rewards or penalties, refining its strategy over time.
Architecture
Signals flow from an input layer, through hidden layers that weigh and combine them, to an output. Training adjusts the weights until the answers improve.
Looking ahead
Train your first model to read handwritten numbers — and see why it works.
Teach a system to tell positive reviews from negative ones.
Investigate where these systems fail — and who is accountable.