Date: Mon, 20 Mar 2017 22:38:45 +0100
Subject: [PATCH] Minor edits
---
slides/talks/2017-1-EDBT-Inference/index.html | 58 ++++++++++---------
1 file changed, 30 insertions(+), 28 deletions(-)
diff --git a/slides/talks/2017-1-EDBT-Inference/index.html b/slides/talks/2017-1-EDBT-Inference/index.html
index e7fdeb71..9e9e645c 100644
--- a/slides/talks/2017-1-EDBT-Inference/index.html
+++ b/slides/talks/2017-1-EDBT-Inference/index.html
@@ -71,11 +71,11 @@
-
+
Disclaimer
Ying could not be here today. If you like her ideas, get in touch with her.
- (she's on the job market)
(If you don't, blame my presentation)
+ (Also, Ying is on the job market)
@@ -93,9 +93,9 @@
Graphical Models
Joint probability distributions are expensive to store
$$p(D, I, G, S, J)$$
- Bayes rule lets us break apart the distribution
+
Bayes rule lets us break apart the distribution
$$= p(D, I, G, S) \cdot p(J | D, I, G, S)$$
- And conditional independence lets us further simplify
+
And conditional independence lets us further simplify
$$= p(D, I, G, S) \cdot p(J | G, S)$$
This is basis for a type of graphical model called a "Bayes Net"
@@ -136,6 +136,7 @@
$\cdot\;0.95$
$\cdot\;0.25$
$\cdot\;0.8$
+ $=\;0.0665$