Skip to content
ml-recall

#hyperparameters

13 questions from Optimization & Calculus, ML Fundamentals, Nearest Neighbors & Kernel Methods, Support Vector Machines, Decision Trees & Ensembles, Probabilistic & Graphical Models, Dimensionality Reduction, Language Modeling

Timed test

13 questions · 13 min · score and explanations at the end

Take a test
Q 1 / 13Score 0 / 0

Question 1: Short answer

mediumAI-checkedOptimization & Calculus

An RMSProp implementation uses s←γs+(1−γ)g2s \leftarrow \gamma s + (1 - \gamma)g^2 and x←x−η g/s+ϵx \leftarrow x - \eta\, g/\sqrt{s + \epsilon}, with ss initialized to 0, ϵ=10−6\epsilon = 10^{-6} and η=0.4\eta = 0.4. If you set γ=1\gamma = 1, what is the effective learning rate that multiplies gg at every step?

a number