University COVID-19 update

The University of Waterloo is constantly updating our most Frequently Asked Questions.

Questions about buildings and services? Visit the list of Modified Services.

Please note: The University of Waterloo is closed for all events until further notice.

Department seminar by Jason Leoppky, University of British ColumbiaExport this event to calendar

Thursday, January 14, 2016 — 4:00 PM EST

A Different Perspective on Statistical Invariance

It is well know that maximum likelihood estimation is invariant to the parameterization of the model.  When the mapping from one parametrization to another is injective this leads to an equivalence of the likelihood and in a Bayesian context an equivalence of the posterior for carefully chosen prior distributions.  The majority of modern statistical procedures requires  numerically sampling the posterior or numerically optimizing the likelihood.  In the context of a Gaussian process model we show that the parameterization of the model is in fact critical to achieving meaningful results and ensuring convergence of the optimizer or the Markov chain.  The lack of numerical invariance is caused by the behaviour of the gradient near the optimum.  In such cases the behaviour of the gradient is crucial in deciding on an appropriate parameterization of the model.

Location 
M3 - Mathematics 3
3127

,

Event tags 

S M T W T F S
25
26
27
28
29
30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1
2
3
4
5
  1. 2021 (51)
    1. October (1)
    2. September (2)
    3. July (1)
    4. June (2)
    5. May (6)
    6. April (8)
    7. March (13)
    8. February (7)
    9. January (12)
  2. 2020 (71)
    1. December (2)
    2. November (13)
    3. October (16)
    4. September (7)
    5. August (5)
    6. July (3)
    7. June (2)
    8. May (1)
    9. March (4)
    10. February (4)
    11. January (14)
  3. 2019 (65)
  4. 2018 (44)
  5. 2017 (55)
  6. 2016 (44)
  7. 2015 (38)
  8. 2014 (44)
  9. 2013 (46)
  10. 2012 (44)