en-de
en-es
en-fr
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
Beyond Relevance
Published on Jul 18, 20113601 Views
Finding similar and relevant media content given a user query or sample image has been at the core of the multimedia retrieval community for a long time. In this talk, I will identify and address mult
Related categories
Chapter list
Faceted Search00:00
Beyond Relevance! - 100:00
Beyond Relevance?00:40
Click-feedback collected01:04
Beyond Relevance! - 201:47
Gain02:20
What to expect?02:51
Overview03:25
CTR03:45
People03:55
Facets in Image Search - 104:14
Understanding Users - 105:53
Predicting Flickr Favorites05:54
Understanding Users - 206:19
Incoming stream of previously unseen candidate images06:41
Facets in Image Search - 206:50
GridFaces - 107:43
Distribution of favorites08:29
Features09:19
GridFaces - 210:37
Learning10:49
Approach11:02
What users search for...11:52
Evaluation (random context) - 112:05
Breakdown of entertainment13:41
Evaluation (random context) - 213:58
Extracting Facets14:39
User session analysis16:25
Extracting Features - 116:35
Pesonalization in Social Media16:44
Classification of user actions18:44
Extracting Features - 218:53
Extracting Features - 320:34
Session 120:59
User Click Feedback22:01
Session 223:48
Evaluation – Overall Performance24:23
User session trends25:11
Evaluation – DCG@p26:22
What do they click on?26:52
Evaluation – nDCG@p26:55
Evaluation – per query26:58
Cond. CTR @ position27:44
Evaluation – Feature Importance28:58
Conditional CTR30:07
Conclusions – Ongoing work31:44
Query31:59
Automated Slideshows33:03
Smarter Thumbnails33:21
Motivation33:30
Understanding users?33:55
Example - 434:18
Slideshow generation34:26
Example - 534:26
Objective34:56
Diversity35:00
Machine learned thumbnail generation35:19
What Properties?35:20
Practical issues35:59
Dimensions of diversity35:59
Cricket example36:45
San Francisco example37:07
Approach (1) 37:31
Detecting and resolving tag ambiguity37:53
Topic Discovery38:04
Example - 638:08
Selecting representative images38:46
Resolving tag ambiguity38:49
Categories and coverage38:58
Sparse Models - 139:32
Sparse Models - 239:33
Performance metric39:36
Nature of ambiguity39:50
Import Vector Machines (IVM) - 139:57
Import Vector Machines (IVM) - 240:02
Import Vector Machines (IVM) - 340:02
Import Vector Machines (IVM) - 440:03
Import Vector Machines (IVM) - 540:04
Import Vector Machines (IVM) - 640:04
Multi – Category IVM - 140:13
Multi – Category IVM - 240:15
Multi – Category IVM - 340:17
Multi – Category IVM - 440:18
Multi – Category IVM - 540:19
Multi – Category IVM - 640:28
Multi – Category IVM - 740:29
Multi – Category IVM - 840:31
Feature space - 140:31
Multi – Category IVM - 940:31
Multi – Category IVM - 1040:32
Multi – Category IVM - 1140:33
Multi – Category IVM - 1240:34
Multi – Category IVM - 1340:34
Multi – Category IVM - 1440:40
LDA on textual meta data of images40:41
Toy Example40:43
Cricket - Results41:04
Flickr tag co-occurrence41:16
Distribution Constraints - 141:21
Feature space - 241:36
Distribution Constraints - 241:43
Distribution Constraints - 341:43
Distribution Constraints - 441:44
Distribution Constraints - 541:44
Distribution Constraints - 641:45
Distribution Constraints - 741:45
Distribution Constraints - 841:46
Distribution Constraints - 941:46
Oil Spill - Uniform Distribution41:47
Feature space - 341:50
Oil Spill - Political view41:56
Oil Spill - Environmental view42:21
Machine Learning Setup42:36
Evaluation?42:49
Results - 243:18
Weighted KL divergence43:28
Summary44:03
Results - 344:33
Probabilistic framework44:40
Performance per category45:22
Performance46:29
Optimization46:59
Summary47:01
Explore-exploit47:41
Example - 748:02
Ambiguity explained48:23
Example - 848:24
Explore/Exploit for Web Content Optimization49:35
Visual Diversity49:59
Need for visual diversity50:15
Towards visual diversification - 151:08
Explore/exploit for images51:38
Towards visual diversification - 255:56
Visual characteristics - 156:37
Visual characteristics - 256:59
Notation (image clustering)58:00
Folding59:07
Reciprocal election01:01:22
Evaluation01:06:13
Results for query: wembley stadium01:08:09
Evaluation criteria01:09:08
Results01:10:23
Ambiguous vs unambiguous - FM index01:12:34
Freshness, Trending, Facets, Slideshows, and Explore01:13:54
Image Fingerprinting01:14:44
DCT fingerprint01:15:03
Example - 101:17:05
Example - 201:18:08
Example - 301:18:19
Multi-dimensional Discrete Cosine Transform - 101:18:49
Multi-dimensional Discrete Cosine Transform - 201:19:42
Multi-dimensional Discrete Cosine Transform - 301:20:22
Alternative methods01:21:48