Suppose we observe the growth (g) of some novel bacteria over time (t). We collect n...

 

Transcribed Image Text:

Suppose we observe the growth (g) of some novel bacteria over time (t). We collect n pairs of data, {(t,8), (t, g),..., (tn, gn)} and decide to model growth as = 0 sin t. We de decide to use squared loss as our loss function. Assume that we are not using regularization. Suppose we are provided with the following data. sin ti gi 1.2 2.5 3.2 -1.4 your answer should contain is 0. No need to simplify.) -0.2 0.1 a) What is the average loss for our model on this dataset, as a function of 0? (The only variable the specific set of values provided in part a.) 0.4 0.6 b) Write an expression for the average loss of our model, R(0), for any set of data with n d) Is this a linear model? Answer "yes" or "no". observations. (Your answer should be in terms of 0, gi, t;, and n, and should not involve . c) What is the value of 0 that minimizes R(0) that you determined in part b? Suppose we observe the growth (g) of some novel bacteria over time (t). We collect n pairs of data, {(t,8), (t, g),..., (tn, gn)} and decide to model growth as = 0 sin t. We de decide to use squared loss as our loss function. Assume that we are not using regularization. Suppose we are provided with the following data. sin ti gi 1.2 2.5 3.2 -1.4 your answer should contain is 0. No need to simplify.) -0.2 0.1 a) What is the average loss for our model on this dataset, as a function of 0? (The only variable the specific set of values provided in part a.) 0.4 0.6 b) Write an expression for the average loss of our model, R(0), for any set of data with n d) Is this a linear model? Answer "yes" or "no". observations. (Your answer should be in terms of 0, gi, t;, and n, and should not involve . c) What is the value of 0 that minimizes R(0) that you determined in part b?