You are on page 1of 3

Subject: INTRODUCTION TO MACHINE LEARNING AND TENSORFLOW

Name: Nguyễn Cao Đạt

Code: HE140197 Classs: SE1509-AI

I. Quiz
Number of question: 5

Multiple choice question:

QN=1 What is MNIST dataset?


a. Which is a set of 70,000 small images of digits handwritten by high
school students and employees of the US Census Bureau.
b. Which is a set of 70,100 small images of digits handwritten by high
school students and employees of the US Census Bureau.
c. Which is a set of 80,000 small images of digits handwritten by high
school students and employees of the US Census Bureau.
d. Which is a set of 60,000 small images of digits handwritten by high
school students and employees of the US Census Bureau.
e.
f.
ANSWER: B

QN=2 How do you calculate Recall in TensorFlow?


a. Use tf.keras.metrics.Recall
b. Use tf.keras.metrics.recall
c. Use tf.keras.Recall
d. Use tf.keras.metrics.recal
e.
f.
ANSWER: A

QN=3 What is batch learning?


a. In batch learning, the system is incapable of learning incrementally: it
must be trained using all the available data.
b. Batch learning will generally take a lot of time.
c. Batch learning will generally take a lot of computing resources.
d. Batch learning is typically done offline.
e. All of the mentioned.
f.
ANSWER: E

QN=4 How do you know about instance-based learning?


a. For this type of learning, system learns the examples by heart, then
generalizes to new cases by comparing them to the learned examples (or
a subset of them), using a similarity measure.
b. For this type of learning, system learns the examples by heart, then
creates new cases by comparing them to the learned examples (or a
subset of them), using a similarity measure.
c. For this type of learning, system learns the examples by heart, then
generalizes to old cases by comparing them to the learned examples (or a
subset of them), using a similarity measure.
d. For this type of learning, system learns the examples by heart, then
generalizes to new cases by comparing them to the learned examples (or
a subset of them), using a different measure.
e.
f.
ANSWER: A

QN=5 How does scientist address underfitting data in machine learning?


a. Selecting a more powerful model, with more parameters.
b. Feeding better features to the learning algorithm (feature engineering).
c. Reducing the constraints on the model (e.g., reducing the regularization
hyperparameter).
d. All of the mentioned.
e.
f.
ANSWER: D

II. Lập trình


 thu thập 20 ảnh xe máy, 20 ảnh xe đạp
 xây dựng mạng CNN phân loại xe máy, xe đạp
 thực hiện augmentation, so sánh kết quả
 lấy các đồ thị từ tensor board

You might also like