video thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Private traits and attributes are predictable from digital records of human behavior

Published on 2013-11-072885 Views

We show that easily accessible digital records of behavior, Facebook Likes, can be used to automatically and accurately predict a range of highly sensitive personal attributes including: sexual orient

Related categories

Presentation

The Human Manifold00:00
The Digital Traces: Online Data00:34
High-Dimensional Observation Space - 101:01
High-Dimensional Observation Space - 201:25
Mapping the Human Manifold01:58
Big Five Personality traits03:46
A Treasure Chest of Data: MyPersonality05:06
The Magic of Machine Learning - 106:46
High-Dimensional Observation Space - 307:26
The Magic of Machine Learning - 207:40
Mapping the Manifold08:08
The Magic of Machine Learning - 309:11
Making Predictions10:07
Demo - 110:37
Prediction Accuracy: Binary variables12:07
Prediction Accuracy: Numeric Variables13:55
Demo - 216:17
What is there to like? Intelligence17:27
Which Likes? Happiness21:11
Which Likes? Extraversion22:36
Press activity - 123:10
Press activity - 223:15
Press activity - 323:17
Press activity - 423:20
Press activity - 523:31
Questions for User Privacy23:47
Conclusions25:56