online algorithm seminar | week 8
For the suppose-to-be-continuous series of online algorithm’s seminar course note (see a pervious note here), here’s a latest one. Today’s theme is introduction to online learning. the so-called “expert setting” There is a decision maker who makes decisions over time horizon $t = 1, 2, \ldots, T$. We expect $T\to \infty$ to be asymptotic in our analysis. There is a set of “actions” — $\lbrace L, H\rbrace$ (assume two for now)....