Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Random Walk with Restart for Automatic Playlist Continuation and Query-specific Adaptations

16 views

Published on

These are the slides for my presentation at the ACM RecSys 2018 conference. We took part in the RecSys Challenge 2018 as Team Radboud.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Random Walk with Restart for Automatic Playlist Continuation and Query-specific Adaptations

  1. 1. Random Walk with Restart for Automatic Playlist Continuation and Query-specific Adaptations Team Radboud: Timo van Niedek timo@mediadistillery.com Arjen P. de Vries arjen@acm.org RecSys ‘18, Vancouver, Canada
  2. 2. Random Walk with Restart for Automatic Playlist Continuation and Query-specific Adaptations Team Radboud: Timo van Niedek timo@mediadistillery.com Arjen P. de Vries arjen@acm.org RecSys ‘18, Vancouver, Canada
  3. 3. Current Challenges of Automatic Playlist Continuation 1)Scalability 2)Playlist purpose 3)Popularity bias
  4. 4. 1. Scalability • RecSys Challenge: 1M playlists, > 2.3M unique tracks
  5. 5. Recent work: Pixie1 • Recommender system at Pinterest • Highly scalable: 3+ billion items, 200+ million users 1) Eksombatchai, C., Jindal, P., Liu, J. Z., Liu, Y., Sharma, R., Sugnet, C., ... & Leskovec, J. (2018, April). Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time. In Proceedings of the 2018 World Wide Web Conference on World Wide Web (pp. 1775-1784). International World Wide Web Conferences Steering Committee.
  6. 6. The Playlist Graph • Neighbor sampling in constant time wrt. number of users & items • Assumption: playlists contain similar tracks (user judgment) • Nearby tracks in graph are similar Timo van Niedek. 2018. Random Walk with Restart for Automatic Playlist Continuation and Query-Specific Adaptations. Master’s thesis. Radboud University, Nijmegen.
  7. 7. Random Walk with Restart • Query track → Random Playlist → Random Track → … • Restart probability: probability of restarting after every step • Track-level recommendations: Sorted list of visit counts • Only local neighborhood is needed: scalable! Timo van Niedek. 2018. Random Walk with Restart for Automatic Playlist Continuation and Query-Specific Adaptations. Master’s thesis. Radboud University, Nijmegen.
  8. 8. Multiple Random Walks with Restart • Playlist-level recommendations: Repeat Random Walk with Restart for every track in playlist Timo van Niedek. 2018. Random Walk with Restart for Automatic Playlist Continuation and Query-Specific Adaptations. Master’s thesis. Radboud University, Nijmegen.
  9. 9. Evaluation • Validation set: sample of the Million Playlist Dataset • 10.000 playlists following structure of challenge set
  10. 10. Restart Probability
  11. 11. 2. Playlist Purpose • Playlist purpose = a set of target characteristics • High quality recommendations match these characteristics • Prefiltering: remove playlists that do not match target characteristics
  12. 12. Measuring Target Characteristics • Target characteristics can be measured from the query playlist • Playlist Title • Query Track Features
  13. 13. Prefiltering approaches • Title-based: remove playlists whose titles do not match query playlist • Feature-based: prune graph based on distance between playlist models using features from Spotify API: Acousticness, danceability, count, energy, explicitness, instrumentalness, liveness, speechiness, tempo, valence, release year
  14. 14. Playlist Model • Estimated from empirical distribution as a histogram
  15. 15. Degree Pruning • Eksombatchai et al. (2018)1 have shown that removing low- quality edges of popular pins is beneficial • Highly popular songs are often added to playlists where they do not belong • Pruning edges of popular songs reduces random walk drift 1) Eksombatchai, C., Jindal, P., Liu, J. Z., Liu, Y., Sharma, R., Sugnet, C., ... & Leskovec, J. (2018, April). Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time. In Proceedings of the 2018 World Wide Web Conference on World Wide Web (pp. 1775-1784). International World Wide Web Conferences Steering Committee.
  16. 16. Degree Pruning • The degree of all track nodes t is updated to by removing the lowest-quality edges according to where t = collection of track features
  17. 17. Hybrid Degree Pruning
  18. 18. Results
  19. 19. 3. Popularity Bias • Major issue in many academic approaches1 1) D. Jannach, I. Kamehkhosh, and G. Bonnin, “Biases in automated music playlist generation: A comparison of next-track recommending techniques,” in Proceedings of the 2016 Conference on User Modeling Adaptation and Personalization. ACM, 2016, pp. 281–285.
  20. 20. 3. Popularity Bias • Major issue in many academic approaches1 • Our methods reduce popularity bias 1) D. Jannach, I. Kamehkhosh, and G. Bonnin, “Biases in automated music playlist generation: A comparison of next-track recommending techniques,” in Proceedings of the 2016 Conference on User Modeling Adaptation and Personalization. ACM, 2016, pp. 281–285.
  21. 21. Summary • Scalable to large graphs • Flexibility of integrating evidence • Reduction of popularity bias

×