Towards Holistic Concept Representations: Embedding Relational Knowledge, Visual Attributes, and Distributional Word Semantics thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Towards Holistic Concept Representations: Embedding Relational Knowledge, Visual Attributes, and Distributional Word Semantics

Published on Nov 28, 20171506 Views

Knowledge Graphs (KGs) effectively capture explicit relational knowledge about individual entities. However, visual attributes of those entities, like their shape and color and pragmatic aspects conce

Related categories

Chapter list

Towards Holistic Concept Representations: Embedding Relational Knowledge, Visual Attributes, and Distributional Word Semantics00:00
KG Embedding Approaches – Overview00:14
Latent Distance Models – TransE – Model01:32
Latent Distance Models – TransE – Model - 101:59
Latent Distance Models – TransE – Model - 102:19
Motivation03:34
Motivation - 104:15
Motivation - 204:30
Visual Embedding – Inception V304:48
Text Embedding – Word2Vec05:15
Text Embedding – Word2Vec - 105:24
Text Embedding – Word2Vec - 205:41
Shared Concept Space06:10
Shared Concept Space - 106:34
Shared Concept Space - 206:45
Shared Concept Space - 307:00
Fusion techniques07:11
Fusion techniques - 107:35
Fusion techniques - 208:00
Empirical Analysis – Word Similarity08:13
Empricial Analysis – Word Similarity – Rank Correlation08:59
Empricial Analysis – Word Similarity – Rank Correlation - 110:01
Empirical Analysis – Word Similarity – Influence of Modalities10:14
Empirical Analysis – Entity Segmentation10:50
Empirical Analysis – Entity Segmentation - 111:20
Empirical Analysis – Entity-Type Prediction11:32
Empirical Analysis – Entity-Type Prediction - 111:47
Empirical Analysis – Entity-Type Prediction - 212:24
Empirical Analysis – Entity-Type Prediction - 312:51
Empirical Analysis – Entity-Type Prediction - 413:40
Future challenges14:44