0.25
0.5
0.75
1.25
1.5
1.75
2
Measuring Semantic Coherence of a Conversation
Published on Nov 22, 20183173 Views
Conversational systems have become increasingly popular as a way for humans to interact with computers. To be able to provide intelligent responses, conversational systems must correctly model the str
Related categories
Chapter list
Measuring Semantic Coherence of a Conversation00:00
Semantic Coherence?00:02
Semantic Coherence ~ Contextual Glue00:57
Semantic Coherence IsA Classification Task01:38
Applications02:59
Contributions03:43
Benchmark04:02
Subgraph induction: Top-k Shortest Paths04:56
Benchmark: Incoherent Dialogues - 105:51
Benchmark: Incoherent Dialogues - 206:06
Subgraph Induction: Performance Bottleneck07:24
Embeddings Classification Approach08:37
Results: Word Embeddings perform Best09:39
Results: KG Embeddings Classification10:09
Results:Word Embeddings Classification10:47
Results: Permutation IsA Difficult Task10:53
Coherence Patterns - 111:32
Coherence Patterns - 212:11
Horizontal Split - 112:41
Horizontal Split - 212:54
Horizontal Split - 312:58
Coherence Patterns13:05
Future Work13:22
Acknowledgements14:24