Copy responder link
Copy
Published
Section 1 of 1
Day 9 · Thu 17 Sep · Singular Learning Theory: Daily Quiz
Day 9 · Thu 17 Sep · Singular Learning Theory (B.3) · Kai Ogden, Matthew Farrugia-Roberts

This is a simple quiz to help you recall what you have learned today, and to confirm you have achieved the learning outcomes. The questions are deliberately easy. Passing this quiz is *not* a replacement for being in the course. You do not need to submit the quiz, and we will not grade it.

Nothing is collected about who you are. If you do submit, your score and the correct answers appear immediately.
A parameter-function map is “degenerate” at w when:
A parameter-function map is “degenerate” at w when:
Question Type
Loading image…
The function it implements is not differentiable at w
There is a nonzero direction in parameter space along which the directional derivative of the parameter-function map vanishes - its Jacobian is rank-deficient
It takes the value zero at w
The loss at w is larger than at every neighbouring point
Add option
or
add "Other"
…
Answer key
(1 point)
Loading...
Loading…
Which statement about the local learning coefficient (LLC) and the loss Hessian at a local minimum is correct?
Which statement about the local learning coefficient (LLC) and the loss Hessian at a local minimum is correct?
Question Type
Loading image…
The LLC always equals the rank of the Hessian
The LLC can distinguish geometries that the Hessian rank cannot
The Hessian rank fully determines the LLC
The LLC is only defined when the Hessian is positive definite
Add option
or
add "Other"
…
Answer key
(1 point)
Loading...
Loading…
If the Hessian of the population loss at a minimum is positive definite, the LLC equals:
If the Hessian of the population loss at a minimum is positive definite, the LLC equals:
Question Type
Loading image…
d, the dimension of the parameter space
The rank of the Hessian
d/2, where d is the dimension of the parameter space
Zero
The largest eigenvalue of the Hessian
Add option
or
add "Other"
…
Answer key
(1 point)
Loading...
Loading…
For a local minimum w* of the population loss in a d-dimensional parameter space, the LLC satisfies:
For a local minimum w* of the population loss in a d-dimensional parameter space, the LLC satisfies:
Question Type
Loading image…
It is at most d/2, with equality exactly in the regular case
It is always exactly d/2
It is at least d/2
It can be any real number, positive or negative
Add option
or
add "Other"
…
Answer key
(1 point)
Loading...
Loading…
Link to Sheets
Insights
Total points distribution
Loading...
Loading responses…
Settings
Release grades
Respondent settings
Global quiz defaults
Default question point value
Point values for every new question
points
Responses
Manage how responses are collected and protected
Presentation
Manage how the form and responses are presented
Defaults
Form defaults
Settings applied to this form and new forms
Question defaults
Settings applied to all new questions
.