Description

1. [4 points] Intro to Machine Learning
Consider the task of classifying an image as one of a set of objects. Suppose we use a convolutional neural network to do so (you will learn what this is later in the semester).
(a) For this setup, what is the data (often referred to as x(i))?
Your answer: The data will be the image to be classfied.
(b) For this setup, what is the label (often referred to as y(i))?
Your answer: The label will be the class which the image belongs to.
(c) For this setup, what is the model?
Your answer: The model we used is Convolutional Neural Network.
(d) What is the distinction between inference and learning for this task?
Your answer: Learning is associated with parameter estimation. After ”learning”, some parameters will be optimized. Inference is thought of as making some sort of prediction directly.
2. [8 points] K-Nearest Neighbors
(1)
x1 x2 y
1 1 2
0.4 5.2 1
−2.8 −1.1 2
3.2 1.4 1
−1.3 3.2 1
−3 3.1 2
(a) Classify each of the following points using the Nearest Neighbor rule (i.e. K = 1) with the squared euclidean distance metric.
Your answer: x1 x2 y
−2.6
1.4 6.6
1.6 1
2
−2.5 1.2 2
(b) Classify each of the following points using the 3-Nearest Neighbor rule with the squared euclidean distance metric.
Your answer: x1 x2 y
−2.6
1.4 6.6
1.6 1
1
−2.5 1.2 2
(c) Given a dataset containing n points, what is the outcome of classifying any additional point using the n-Nearest Neighbors algorithm?
Your answer: In this case k = n, the smoothing effect decreased variance which will effect overall performance negatively. It will result in underfitting.
(d) How many parameters are learned when applying K-nearest neighbors?
2
Your answer: None. There are zero parameters being learned when applying Knearest neighbors.
3