Monday: Coding
x = 5
y = 6
x = x + y
Find x
What is 11?
The clustering technique uses the means of surrounding data points to classify them into groups
What is K-clustering/K-means clustering?
These are the middle layers in a neural network
What are hidden layers?
The matrix that convolves with the original image to detect features
What is a filter?
This is the method used to find the steepest descent in the error function
What is a gradient?
This operator gives the remainder of a / b
What is the modulo operator?
K-clustering is a type of _____ learning
What is unsupervised?
These are the adjustable coefficients that we train during the "training" phase of the neural network
What are weights/biases?
This process simplifies convolved images so that they take up less memory
What is pooling?
This is the value that determines how much each weight is changed by for each iteration of the neural network training
What is learning rate?
These characters surround text in code to make them into strings
What are double quotation marks?
This is the method in which the first K-means are initially chosen
What is random?
The function used to normalize neurons to numbers between 0 and 1
What is an activation function?
Before putting a picture into a vanilla neural network, it must undergo this process
This is the term for finding the direction and magnitude of the steepest descent, and moving a step in that direction
What is back-propogation?
The result of 9/5
What is 1?
This is the r value of a perfect negative correlation
What is -1?
The output of a neural network relative to the inputs
What is a weighted average?
This activation function helps normalize output values to turn them into probabilities
What is softmax?
This optimization method uses an estimate of the gradient to decide which direction to move in
What is stochastic gradient descent?
The following Python code loops forever. What is it missing?
n = 10
while n<100:
print n
What is an increment statement?
This loss function is used in regression analysis
What is mean squared error?
The training has an accuracy of 99% but the testing has an accuracy of 54%
What is overfitting?
Our head instructors name
Who is Millard McElwee?
The following issue is noticed:
The machine learning algorithm is taking forever. On closer inspection, we notice that the weights move so far that they keep going past the minima and never actually settle into it. How can we fix this?
What is decrease the learning rate?