CS 2541: Course Survey Machine Learning for Health

CS 2541 will give a broad overview of machine learning for health.

We discuss the recent successes of of graphical models, deep learning, time-series analysis, and transfer learning in the context of health.

Students will choose and complete a course project, and make project presentations at the end of the course.

We require that students have the appropriate background based on several criteria to foster an interesting and interactive course. We are particularly interested in students with a passionate interest in pursuing a research project related to natural language, and also students prepared for the challenging programming aspects of the course.

This course requires a strong background in linear algebra and probability theory, or strong grades in the machine learning course. Familiarity with programming and software engineering is beneficial, but not required.

Email address *

Name *

Your answer

UTOR ID *

Your answer

Education Level *

Master

PhD

Other:

Related ML Course Work *

CSC321

CSC411

CSC412

STA414

ECE521

Other:

Required

Previous Experience *

Extensive Python Coding

Healthcare Background

Matlab/Numpy

Tensorflow/PyTorch

Statistical Modeling (e.g. R, Stan)

Other:

Required

Why are you interested in taking this class? (Please provide a longer substantive response, we will read these closely) *

Your answer

Possible Project Ideas? *

Your answer

Consider the following generative process describing the joint distribution p(Z,X): Z ~ Bernoulli(0.2), X ~ Gaussian(mu_Z, 0.5), where mu_0=3 and mu_1=5. Which of the following plots is the marginal distribution p(X)? *

A

B

C

D

Gradient descent will always converge to a global minimum, even for non-convex functions. *

True

False

Stochastic (sub)gradient descent can be used to learn a Support Vector Machine (SVM) classifier. *

True

False

You go for your annual checkup and have several lab tests performed. A week later your doctor calls you and says she has good and bad news. The bad news is that you tested positive for a marker of a serious disease, and that the test is 97% accurate (i.e. the probability of testing positive given that you have the disease is 0.97, as is the probability of testing negative given that you don’t have the disease). The good news is that this is a rare disease, striking only 1 in 20,000 people. What are the chances that you actually have the disease? *

1/20000 = 0.00005

0.97*0.00005 ≈ 0.00005

0.03*0.00005 = 0.0000015

0.97*0.00005 / (0.97*0.00005 + 0.03*0.99995) ≈ 0.0016

What would be the value of output z in the neural network below for the input x1=1 and x2=0 ? *

-3

-2

-1

0

1

2

3