TODOs

[shimingyoung]

1. check the gradient_log_loss (Rajat)
2. check the L (Rajat, Shiming)
3. finalize the dirty_model_logistic function (Shiming, Rajat)
4. Write a main script to run the dirty model (Shiming)

[rajathpatel23]

Dear Dr. yang, I have worked don the dirty model logistic function and uploaded my version of the code. Please do check. I will soon try and upload your version of dirty model logistic code and update.

Thanks, Rajat Patel

[rajathpatel23]

@shimingyoung Hello Dr. Yang, I have written the function for L as given in the FISTA algorithm paper, The function is as follows: L = max Eigen value of A^T A function is as follow as:

def Lipchitz_function(input_matrix): aT = np.transpose(input_matrix) product = np.dot(aT, input_matrix) w, v = np.linalg.eig(product) print(v.shape) print(w) return(2* max(w))

def adaptive_Lipchitz(old_L, eta): L_new = eta*old_L return(L_new)