0.25
0.5
0.75
1.25
1.5
1.75
2
Harnessing Diversity in Crowds and Machines for Better NER performance
Published on Jul 10, 2017783 Views
Over the last years, information extraction tools have gained a great popularity and brought significant performance improvement in extracting meaning from structured or unstructured data. For example
Related categories
Chapter list
Harnessing Diversity in Crowds and Machines for Better NER Performance00:00
What’s wrong with the ground truth for NER?00:19
What’s wrong with the ground truth for NER? - 101:08
What’s wrong with the ground truth for NER? - 202:06
What’s wrong with the ground truth for NER? - 303:02
… but, then ...03:37
How do state-of-the-art NER tools perform on this ground truth?03:55
… let’s see how NER tools perform on this GT 05:25
… and now let’s see their overall performance 06:14
… and now let’s see their overall performance - 107:17
Difficult to understand the reliability of different NER tools 07:39
Hypothesis07:53
Multi-NER vs. SOTA Single-NER08:19
Multi-NER vs. SOTA Single-NER - 109:00
Multi-NER vs. SOTA Single-NER - 209:23
Multi-NER vs. SOTA Single-NER - 310:10
Disagreement Aware Multi-NER vs. Single-NER - 111:24
Hypothesis - 111:49
Crowdsourcing for better ground truth12:04
Crowdsourcing task12:26
Crowdsourcing for better NER 13:25
Crowdsourcing for better ground truth -1 15:52
Conclusions & future work16:30