Introduction to Embedded Machine Learning Week 2 Quiz Answers

Question 1) The activation function allows a perceptron (node) to act as a linear regression function.

Question 2) A single node in a neural network performs the following actions in order:

Multiplies each input value by a weight

Sums the products of step 1

Adds a bias term to the sum computed in step 2

Applies an activation function to the sum from step 3 to produce a single value output

Question 3) “Backpropagation” is the process of computing a prediction given an input sample and comparing the model’s prediction to that sample’s known label to calculate a loss value.

Question 4) Model parameters (weights and bias terms) are updated manually by the programmer during training.

Question 5) Performing forward and backward passes on the model with the entire data set is called what?

Evaluation, Underfitting, and Overfitting Quiz Answers

Question 1) What is the total accuracy given the following confusion matrix? Please express your answer as a decimal rounded to 3 decimal places (For example, 68.23% would be 0.682).

Question 2) What is the total number of true positives given the following confusion matrix? Your answer should be a positive, whole number (e.g. 123).

Remember that “true positive” is where the predicted label matches the actual label.

Question 3) What is the F1 Score of class C given the following confusion matrix? Please express your answer as a decimal rounded to 3 decimal places (For example, 68.26% would be 0.683).

The F1 score is calculated by the following equation:

Where:

Question 4) You plot the training and validation loss over the training period of the model, and you see the following. What can we conclude about the model?

Question 5) Which of the following are good ways to try to reduce overfitting? Select all that apply.

Question 1) You collect raw data from a 3-axis accelerometer for a period of 1 second at a sample rate of 100 Hz. Your feature extraction method is to compute the root mean square (RMS) of all of the values in each axis. How many dimensions is your feature set for one sample? Your answer should be a whole, positive number.

Question 2) The features you extracted during the training process must have the same number of dimensions as the features you extract during inference.

Question 3) Your machine learning model is a neural network that ends with the softmax function, and you have trained it to predict the label among 5 different classes. How many outputs does your model have?

Question 4) After performing inference, the outputs of the softmax function should always add up to 1.0.

Question 5) After performing inference, your model outputs 5 probabilities, each pertaining to the probability or confidence rating of each class. Finding the lowest probability will give you which class the model is predicting that the data belongs to.

Anomaly Detection Quiz Answers

Question 1) You wish to identify the spoken keyword “Alexa.” An anomaly detection system is a good way to accomplish this task.

Motion Classification and Anomaly Detection Quiz Answers

Question 1) The activation function provides a non-linear output of a single node in a neural network.

Question 2) Hyperparameters are updated automatically during the training process.

Question 3) The process of automatically updating a neural network’s parameters based on the calculated loss between the predicted label and actual label is known as a forward pass.

Question 4) What is the total accuracy given the following confusion matrix? Please express your answer as a decimal rounded to 3 decimal places (For example, 68.23% would be 0.682).

Question 5) What is the F1 Score of class A given the following confusion matrix? Please express your answer as a decimal rounded to 3 decimal places (For example, 68.23% would be 0.682).

The F1 score is calculated by the following equation:

Where:

Question 6) You plot the training and validation accuracy over the training period of the model, and you see the following. What can we conclude about the model?

Question 7) You are creating a motion classification system, and your feature extraction method includes calculating the root mean square (RMS) and the frequency value of the 3 highest peaks of the power spectral density (PSD) of each axis on a 3-axis accelerometer from a 10 millisecond sample period. How many dimensions should the input to your model be?

Question 8) You deploy your machine learning model to an embedded system. The model ends with the softmax function, giving you 4 outputs, one for each class. How do you find the predicted class?

Question 9) The identification of rare events is known as anomaly detection.

Question 10) Gathering more data is one way to reduce overfitting.