Personalized Celebrity Video Search: Cross-space Mining Approach

This study presents a novel approach for personalized celebrity video search based on cross-space mining, aiming to match user interests with celebrity popularity. By learning the interest space of users and the popularity space of celebrities, the framework correlates the two spaces to enhance search results. Motivations, framework, approach, and experiments are detailed, addressing the diverse interests of users and the varying popularity of celebrities across different fields. The proposed solution involves leveraging user-topic modeling, interest space mapping, query re-ranking, and popularit.y space analysis to optimize the search engine experience. Experimental conclusions highlight the effectiveness of the approach in aligning user preferences with celebrity content.

• Celebrity Video Search
• Cross-space Mining
• User Interests
• Popularity Distribution
• Topic Modeling

malo_2 Follow

Uploaded on Oct 06, 2024 | 0 Views

Download Presentation

Please find below an Image/Link to download the presentation.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.

Presentation Transcript

1. Personalized Celebrity Video Search Based on Cross-space Mining Zhengyu Deng, Jitao Sang, Changsheng Xu 1 Institute of Automation, Chinese Academy of Sciences 2 Chinese-Singapore Institute of Digital Media 1

2. Outline Motivation Framework Approach Experiment Conclusions 2

3. Motivation Celebrities are often popular in multiple fields and user interests are diverse. like Sports video User 1 like Entertainment video Beckham User 2 like Interview video User 3 3

4. Motivation Celebrities are often popular in multiple fields and user interests are diverse. Sports video Beckham like Entertainment video like User Bieber like Lady Gaga Music video 4

5. Motivation Non-personalized search David Beckham Daily life Sports Interview 5

6. Motivation Problem and solution Problem Users have different interest distribution and celebrities have different popularity distribution. How to match user interest and celebrity popularity? Solution Learn interest space of users and popularity space of celebrities, then correlate the two spaces. 6

7. Framework Associated tags User Topic modeling Interest space Query Re-rank Map Search engine Popularit y space Topic modeling Celebrity 2024/10/6 7

8. Approach User Interest Space Celebrity Vocabulary Popularity Space U1 C1 X1 W1 Z1 U2 C2 Z2 X2 W2 Um Zp Cn Xq Wx Random walk LDA LDA KL-Divergence P(Zi|Ui) P(Wi|Zi) P(Wi|Ti) P(Ti|Ci) P(Zi|Xi) 8

9. Approach Random walk Vjis the initial probabilistic score; pijis the transition matrix; rk(j) donote the relvence score of node j at iteration k (1) Rewrite as (2) The unique solution (3) 9

10. Approach KL-Divergence ????? ? (?) is from interest (popularity) space. The KL- Divergence between them is ???(? || ?) =1 2( ??(?)??? ? ? ?+ ??(?)???(?) ?(?)) (4) where ?(?) (? ? ) denote the distribution score of topic ? ? on word ?. The similarity ???of topic ? and ? is defined as the inverse of KL-Divergence. ???= 1/???(? || ?) (5) 10

11. Approach Video Projection Given a celebrity video ?? 1, project it to interest space ? ? ?? 1 where K is the topic number of interest space. M is the dimension of the vocabulary. = ? ??? 1 (6) 11

12. Approach Video re-ranking Given a user ? and celebrity ?, the score of ? is ? ????? ?,?,?) = ?=1 ? ??? ? ??? ? ??? = ?=1 ? ??? ? ??? ?=1 where K(L) is the topic number of interest (popularity) space, ??(??) is the ? th (? th) topic of interest (popularity) space, ? ???? is approximated by the inverse of KL-Divergence. (7) ? ? ? ? ??? ? ???? 12

13. Experiments Data Preparation Celebrity list The World's Most Powerful 100 Celebrities List http://www.forbes.com/wealth/celebrities/list The 30 Most Generous Celebrities http://www.forbes.com/sites/andersonantunes/2012/0 1/11/the-30-most-generous-celebrities/3/ Top 200 Sexiest Actor http://www.imdb.com/list/Uun6vT7hWeM/ For each celebrity, 200 videos are downloaded from YouTube. 13

14. Experiments User and Celebrity Profiling User registration info., favorite and uploaded videos raw tags stop words WorldNet noun tags. Celebrity Wikipedia Entry WorldNet noun tags celebrity user total Size 286 200 486 Tags Number 11424 5833 12073 14

15. Experiments Experimental Setting Experiment data 143 users 106 celebrities Experiment setup Each user have some videos related with a specific celebrity. Leave this videos out and learn topics. Rank this celebrity s videos for the user. Evaluation f-Measure 15

16. Experiments Topic simples 16

17. Experiments Doc-Topics distribution E.g. Celebrity Beckham Topic Probability of appearance 7 4 0 8 3 1 6 9 5 2 0.6229086229086229 0.1956241956241956 0.04967824967824968 0.03963963963963964 0.022393822393822392 0.018532818532818532 0.017245817245817245 0.014414414414414415 0.014414414414414415 0.005148005148005148 17

18. Experiments Topic-terms distribution E.g. <topic id= 7"> <word weight="0.018062955825114312" count="478">jay</word> <word weight="0.01726939500434569" count="457">messi</word> <word weight="0.016891508899217776" count="447">real</word> <word weight="0.016551411404602652" count="438">ronaldo</word> <word weight="0.01640025696255149" count="434">kanye</word> <word weight="0.015644484752295656" count="414">west</word> <word weight="0.015606696141782866" count="413">wayne</word> <word weight="0.014964289763065412" count="396">lil</word> <word weight="0.013414956732040963" count="355">hop</word> <word weight="0.013226013679477006" count="350">lionel</word> <word weight="0.01311264784793863" count="347">beckham</word> <word weight="0.01231908702717001" count="326">beyonce</word> <word weight="0.012054566753580472" count="319">cristiano</word> <word weight="0.011941200922042096" count="316">soccer</word> <word weight="0.011941200922042096" count="316">football</word> 18

19. Experiments Topic-terms distribution E.g. <topic id= 4"> <word weight="0.026509629402286503" count="1382">show</word> <word weight="0.014904473260185682" count="777">david</word> <word weight="0.014444103429755236" count="753">ellen</word> <word weight="0.01430982889587969" count="746">tv</word> <word weight="0.012027161819995396" count="627">comedy</word> <word weight="0.01112560423540244" count="580">jennifer</word> <word weight="0.010857055167651347" count="566">interview</word> <word weight="0.010550141947364382" count="550">degeneres</word> <word weight="0.010166500422005677" count="530">funny</word> <word weight="0.009245760761144787" count="482">letterman</word> <word weight="0.008689480549374665" count="453">hollywood</word> <word weight="0.008497659786695312" count="443">late</word> <word weight="0.007979743727461061" count="416">talk</word> <word weight="0.007615284278370291" count="397">celebrity</word> <word weight="0.006943911608992557" count="362">television</word> 19

20. Experiments Different approaches Non-personalized One unified space Our method 0.5 0.4 score @ k Average F- -score @ k 0.3 Average F 0.2 0.1 0 5 10 15 20 25 30 35 40 45 50 k k 20

21. Experiments Impact of random walk 0.35 0.3002 0.3001 0.297 0.296 0.293 0.29 0.288 0.3 0.286 0.281 Average F- -score score 0.25 Average F 0.2 0.15 0.118 0.1 0.067 0.05 0 0 0.1 0.2 0.3 0.4 0.5 ? ? 0.6 0.7 0.8 0.9 1 21

22. Conclusions Conclusions We presented a cross-space mining method to exploit the correlation between user preferences and celebrity popularities. Future work Instead of returning a ranking list, we will try to visualize the search results into semantically consistent groups. Investigate the issue of personalized query understanding in more general personalized search applications. 22

23. Thank you! Q&A? 23