en
0.25
0.5
0.75
1.25
1.5
1.75
2
Large Scale Learning at Twitter
Published on Aug 13, 201224668 Views
Twitter represents a large complex network of users with diverse and continuously evolving interests. Discussions and interactions range from very small to very large groups of people and most of them
Related categories
Chapter list
Large Scale (Machine) Learning at Twitter00:00
Earthquake Los Angeles, July 29, 200803:52
Japan Replies (1)04:51
Japan Replies (2)05:43
What is Twitter06:13
#egypt, #tunesia, #libya, #syria ...07:03
#ocws, #occupywallstreet, ...07:44
The Scale of Twitter08:13
Large scale infrastructure of information delivery08:48
Support for user interaction09:38
Problems we are trying to solve: Relevance10:06
Problems we are trying to solve: Who to follow11:36
Problems we are trying to solve: Content recommendation12:23
(other) problems we are trying to solve12:54
Recommendation/Personalization13:18
Is this BIG data ?13:50
Challenges14:47
What type of machine learning?15:25
ML for social networks (1)16:20
ML for social networks (2)18:15
ML for social networks (2)19:04
Are there wheels not to reinvent?19:22
Analytics Ecosystem20:28
Maximizing the use of Hadoop21:19
AVOID: "janky" analysis of messy data22:17
Leveraging off-line tools23:43
Large scale learning frameworks24:08
Our extensions25:23
Build/reuse/integrate26:28
MapReduce27:06
Java programming28:25
PigML29:14
Training a model in Pig (1)29:35
Training a model in Pig (2)29:53
Training a model in Pig (3)30:05
Applying a model in Pig (1)30:28
Applying a model in Pig (2)30:41
Model training UDF internals31:00
Supervised classification in a nutshell (1)32:08
Supervised classification in a nutshell (2)33:17
Ensembles35:21
Classifier Training / Making Predictions35:48
Ensembles: continued36:59
Further advantages of parallelism37:17
Example: tweet sentiment detection37:27
Diminishing returns ...39:14
Iterative algorithms40:42
Topic modeling42:09
Example: modeling topic distribution42:41
Mahout/PigML integration43:15
LDA applications44:06
Anchoring LDA44:12
ML/Data Mining we contribute to45:12
ML outside of Twitter45:46
Publications mentioning ...46:28
Quick search46:58
Spam/spammer modeling47:17
Example: normal interactions48:59
Example: spammy interactions49:38
ML @ Twitter50:20
Thank you50:46