Embedding Mapping Approaches for Tensor Factorization and Knowledge Graph Modelling thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Embedding Mapping Approaches for Tensor Factorization and Knowledge Graph Modelling

Published on Jul 28, 20161194 Views

Latent embedding models are the basis of state-of-the art statistical solutions for modelling Knowledge Graphs and Recommender Systems. However, to be able to perform predictions for new entities and

Related categories

Chapter list

Embedding Mapping Approaches for Tensor Factorization and Knowledge Graph Modelling00:00
Agenda00:09
Introduction to Latent Embedding Models - 100:43
Introduction to Latent Embedding Models - 203:38
Motivation to Embedding Mapping04:56
Approach 0: Factorization Models with Closed-Form Mappings05:49
Approach 1: Post Mapping - 107:36
Untitled07:56
Approach 1: Post Mapping - 308:18
Approach 2: Hatting Algorithm09:07
Approach 3: Back-Propagation - 109:58
Approach 3: Back-Propagation - 210:37
Experiment 1: MovieLens Data11:53
Experiment 2: FreeBase Data13:31
Experiment 3: Amino Acid Data14:57
Conclusions and Outlooks18:50
(Selected) References21:15
Thank you for your attention21:19