0.25
0.5
0.75
1.25
1.5
1.75
2
Disrupting the semantic comfort zone
Published on Jul 10, 20171202 Views
Ambiguity in interpreting signs is not a new idea, yet the vast majority of research in machine interpretation of signals such as speech, language, images, video, audio, etc., tend to ignore ambiguity
Related categories
Chapter list
Disrupting the semantic confort zone00:00
Personal Semantics00:09
Semantic Social Life01:38
Intelligent educational systems02:08
Intelligent educational systems - 102:37
Intelligent educational systems - 202:48
Intelligent educational systems - 303:13
Digital humanities03:41
User modeling and user-adapted interaction04:39
CrowdTruth team05:13
The evolution of the semantic web05:46
A long time ago in a galaxy far, far away06:48
80’s - empire of the experts07:31
80’s - empire of the experts - 108:10
Rule-based expert systems09:06
90’s - knowledge acquisition from experts11:01
90’s - knowledge acquisition from experts - 111:36
90’s - knowledge acquisition from experts - 211:56
Knowledge engineering and management14:04
00’s - interoperability & standards odyssey15:01
10’s - AI Awakens15:46
201116:40
10’s – Big Data17:07
10’s – Crowds17:45
Untitled17:59
The semantic comfort zone18:37
One truth19:14
Use case: video archive enrichment20:14
Use case: video archive enrichment - 120:42
Use case: video archive enrichment - 221:17
Use case: video archive enrichment - 322:03
Use case: real world QA for Watson22:46
Use case: real world QA for Watson - 122:51
Contradicting evidence23:51
Use case: medical relation extraction for Watson24:37
Antibiotics25:18
Use case: map music to moods26:25
Which is the mood most appropriate for each song?26:42
Semantic comfort zone27:48
Semantic comfort zone - 128:33
Semantic comfort zone - disrupted28:52
It's time to break free29:06
interestingly29:13
Wisdom of crowds29:15
wisdom of crowds - 129:52
WWII Math Rosies30:20
NASA’s computer room30:35
Can we harness it?31:01
CrowdTruth31:56
CrowdTruth - 132:45
CrowdTruth - 234:22
Changes needed video archive enrichment35:28
Crowdsourcing video tagging36:27
Engage crowds through continuous gaming37:13
Engage crowds through continuous gaming - 137:40
Time-based bernhard just “tags”37:50
Objects 38:16
Persons38:27
Locations38:31
User vocabulary38:59
Crowdsourcing medical relation extraction40:15
Does this sentence express TREATS(Antibiotics, Typhus)?41:03
Experts hallucinate41:55
Medical relation extraction42:52
Medical relation extraction44:14
Medical relation extraction - 244:40
Medical relation extraction - 344:43
Learning curves45:45
Learning curves extended47:08
# of workers: Impact on sentence-relation Score48:03
Training a relation extraction classifier48:49
Getting comfortable again49:24
Take home message50:16
Datasets collected and processed with CrowdTruth50:35