Title: Dependent%20Hierarchical%20Normalized%20Random%20Measures%20for%20Dynamic%20Topic%20Modeling%20%20Changyou%20Chen,%20Nan%20Ding%20and%20Wray%20Buntine%20ICML%202012
1Dependent Hierarchical Normalized Random Measures
for Dynamic Topic Modeling Changyou Chen, Nan
Ding and Wray BuntineICML 2012
- Presented by Mingyuan Zhou
- Duke University, ECE
- October 24, 2012
2Introduction
- NRM normalized random measures with independent
increments - Superposition, subsampling and point transition
of NRM - Dependent hierarchical NRM
- Dynamic topic modeling
3Normalized Random Measures
- Poisson process
- Completely random measures (CRM)
4Normalized Random Measures
- Completely random measures (CRM)
5Normalized Random Measures
- Slice sampling NRMs
- Ref Griffin, J.E. and Walker, S.G. Posterior
simulation of normalized random measure mixtures.
J. Comput. Graph. Stat., 2011.
6Normalized Random Measures
- Normalized generalized gamma process
-
7Dynamic topic modeling with dependent
hierarchical NRMs
- Ideas
- Inherit topics from the previous time frame
through three dependency operators - Superposition
- Subsampling
- Point transition
- Generate new topics
-
8Dynamic topic modeling with dependent
hierarchical NRMs
9Dynamic topic modeling with dependent
hierarchical NRMs
10Dynamic topic modeling with dependent
hierarchical NRMs
11Dynamic topic modeling with dependent
hierarchical NRMs
- Properties of the dependence operators
-
12Dynamic topic modeling with dependent
hierarchical NRMs
13Dynamic topic modeling with dependent
hierarchical NRMs
- Original and reformulated model
-
14Sampling
- Sampling under the Chinese restaurant metaphor
-
15Sampling
- Sampling under the Chinese restaurant metaphor
-
16Sampling
17Sampling
18Sampling
19Experiments
20Experiments
21Experiments
22Experiments
23Experiments
24Experiments
25Conclusions