0.25
0.5
0.75
1.25
1.5
1.75
2
Collaborative Learning of Preferences for Recommending Games and Media
Published on Jan 24, 20124424 Views
The talk is motivated by our recent work on a recommender system for games, videos, and music on Microsoft’s Xbox Live Marketplace with over 35M users. I will discuss the challenges associated with su
Related categories
Chapter list
Collaborative Learning of Preferences for Recommending Games and Media00:00
Outline - 100:23
Outline - 201:18
Xbox Live Marketplace01:20
Recommendation of Xbox games02:16
Isn’t this just like the Netflix problem?02:45
Rich User and Item Metadata03:58
Rich User/Item Interaction Data05:00
Games, Play-Count, and Play-Time07:47
Commercial Decision Problem09:39
Large-Scale Systems Challenge11:08
From Interaction to User Preferences13:22
Absolute Ratings vs Preferences15:01
Outline - 317:25
Factor Graphs / Trees17:48
Factor Graphs and Inference19:18
Approximate Message Passing - 121:13
Approximate Message Passing - 222:32
Outline - 423:02
MATCHBOX23:35
Map Features To ‘Trait’ Space23:41
Matchbox With Metadata - 125:14
Matchbox With Metadata - 225:48
Message Passing For Matchbox - 126:56
Message Passing For Matchbox - 227:29
User/Item Trait Space27:42
Outline - 528:46
TRUESKILL28:58
Ranking and Matchmaking29:03
The Skill Rating Problem - 129:33
The Skill Rating Problem - 230:03
The Skill Rating Problem - 330:15
The Skill Rating Problem - 430:18
The Skill Rating Problem - 530:45
The Skill Rating Problem - 630:49
Multiple Team Match Outcome Model - 130:53
Multiple Team Match Outcome Model - 231:24
Efficient Approximate Inference31:44
Applications to Online Gaming - 132:25
Applications to Online Gaming - 232:39
Convergence Speed33:02
Xbox 360 & Halo 333:13
Halo 3 in Action33:35
Halo 3 Analysis: Fair Matches?33:39
Skill Distributions of Online Games33:40
Outline - 634:01
Collaborative Preference Model34:05
Matchbox Generates Raw Ratings34:46
Proposed observation model35:03
Order Constraint Through Rating35:05
Outline - 735:33
Sushi preferences (Kamishima, 2003)35:37
Sushi preference data - 136:08
Sushi preference data - 236:44
Xbox usage data - 136:58
Xbox usage data - 237:46
Xbox usage data37:50
Future work38:35