Agenda and minutes
1. Status updates. RS: He and VJB submitted a presentation slide deck to the WMSCI conf. It will be an "asynchronous" item. He's also giving a 40-minute keynote. Will be absent from here next week due to conflict. PT: Working on the system. DB: discussed some images at https://www.google.com/search?q=global+minimum+surface and https://raw.githubusercontent.com/ritchieng/machine-learning-stanford/master/w1_linear_regression_one_variable/2_params.png from https://www.ritchieng.com/one-variable-linear-regression/. The lesson is that a parameter vector is a hypothesis; multiple local minima mean multiple hypotheses; a shallow minimum suggests a range of plausible parameter values.
2. Literature discussion: Read "6)" of this. We will start on section IV next time.
I run late too late because was at meeting and could not catch up with the rest of meeting today
ReplyDelete.. Sorry