Learning Bayesian Statistics

#78 Exploring MCMC Sampler Algorithms, with Matt D. Hoffman

7 snips
Mar 1, 2023
Ask episode
Chapters
Transcript
Episode notes
1
Introduction
00:00 • 3min
2
Learning Bayesian Statistics - Matt Huffman
02:51 • 4min
3
Late and Dirichlet Allocation - What About You?
06:39 • 4min
4
McMc Methods That Take Advantage of Hardware Acceleration
10:22 • 6min
5
What's the Common Threat to Machine Learning?
16:05 • 2min
6
When Do You Need Fully Bayesian Inference in a Decision Making Model?
17:41 • 2min
7
The Bayesian Framework and Deep Learning
20:01 • 2min
8
Generalized Hamiltonian Monte Carlo
21:56 • 5min
9
Generalized HMC
27:07 • 4min
10
Is Meads Complementary to HMC?
31:13 • 4min
11
Is There a Barrier to Entry?
34:50 • 2min
12
Using a Nested R Hat to Compare Chains
36:50 • 3min
13
Is It Necessary to Tune the Parameters of Your Sampler?
39:50 • 2min
14
The Biggest Herderal in the Bayesian Workflow?
41:32 • 4min
15
Is G HMC a Substitute for HMC?
45:50 • 3min
16
Is There Really a Place for Bayesian Methods in Machine Learning?
48:51 • 4min
17
I'd Like to See a Future Where the Formalism Is Simple and Easy to Communicate
53:05 • 3min
18
Is There a Food Based Name for the Samplers?
55:58 • 2min
19
What's the Biggest Problem in the Show?
57:49 • 2min
20
Earning Patient Stats - Part 2
59:42 • 3min