Machine Learning in Feedback Systems

This page presents lecture materials and calendar for CS 6784-002: Machine Learning in Feedback Systems taught by Sarah Dean at Cornell University in fall 2022.

The most up-to-date material and information can be accessed here (requires Cornell login).

Date	no.	Topic	Materials
8/22	1	Overview & Supervised Learning	Slides, Recording
8/24	2	Supervised Learning: Non-discrimination, least squares	Slides, Recording
8/29	3	Online Learning via Online Convex Optimization	Slides, Recording
8/31	4	Dynamical Systems: equilibria, stability, linear systems	Slides, Recording
9/5		No Meeting: Labor Day
9/7	5	Dynamical Systems: nonlinear stability, inputs, outputs	Slides, Recording
9/12	6	HSNL18 Fairness without demographics in repeated loss minimization
9/14	7	PZMDH20 Performative prediction
9/19	8	State Estimation	Slides, Recording
9/21	9	System Identification	Slides, Recording
9/26	10	Guest Lecture by Max Simchowitz of MST+18 Learning Without Mixing	Recording
9/28	11	TMP20 Sample complexity of Kalman filtering
10/3	12	Contextual Bandits: setting & estimation	Slides, Recording
10/5	13	Contextual Bandits: low regret algorithms	Slides, Recording
10/10		No Meeting: Fall Break
10/12	14	Optimal Control & Dynamic Programming	Slides, Recording
10/17	15	Optimal Linear Control	Slides, Recording
10/19	16	Model-Based RL & Robustness	Slides, Recording
10/24	17	FGKM18 Global convergence of policy gradient methods
10/26	18	Guest Lecture by Stephen Tu: Learning from Many Trajectories	Recording
10/31	19	KTR19 Finite time analysis of approximate policy iteration and IZY22 How to learn when data gradually reacts to your model
11/2	20	Safe Control	Slides, Recording
11/7	21	Model Predictive Control	Slides, Recording
11/9	22	Model Predictive Control	Slides, Recording
11/14	23	RB17 Learning model predictive control for iterative tasks
11/16	24	DSA+20 Fairness is not static
11/21	25	FLD21 Algorithmic fairness and the situated dynamics of justice
11/23		No Meeting: Thanksgiving
11/28	26	JKL+20 Fair prediction with endogeneous behavior or LWH+20 The disparate equilibria of algorithmic decision making
11/30	27	ZCL+21 Towards content provider aware recommender systems
12/5	28	HJMR19 A geometric model of opinion polarization