ACM RecSys 2012: Recommender Systems, Today

739 views
633 views

Published on

Neal Lathia @Cambridge_Uni talk at @ds_dln #strataconf 02/10/12

Published in: Technology
0 Comments
5 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
739
On SlideShare
0
From Embeds
0
Number of Embeds
34
Actions
Shares
0
Downloads
0
Comments
0
Likes
5
Embeds 0
No embeds

No notes for slide

ACM RecSys 2012: Recommender Systems, Today

  1. 1. acm recsys 2012:recommender systems, today@neal_lathia
  2. 2. warning:daunting task lookout for twitter handles
  3. 3. why #recsys? information overloadmailing lists; usenet news (1992) see: @jkonstan, @presnick
  4. 4. why #recsys? information overload filter failuremovies; books; music (~1995)
  5. 5. why #recsys? information overload filter failure creating valueadvertising; engagement; connection (today)
  6. 6. @dtunkelang
  7. 7. (1) collaborative “based on the premise that people looking forinformation should be able to make use of what others have already found and evaluated” (maltz & ehrlick)
  8. 8. (2) query-less“in September 2010 Schmidt said that one day the combination of cloud computing and mobile phones would allow Google to pass on information to users without them even typing in search queries”
  9. 9. (3) discovery engines“we are leaving the age of information and entering the age of recommendation” (anderson)
  10. 10. input: ratings, clicks, views users → items process: SVD, kNN, RBM, etc. f(user, item) → prediction ~ ratingoutput: prediction-ranked recommendations measure: |prediction – rating| 2 (prediction – rating)
  11. 11. traditional problemsaccuracy, scalability, distributed computation, similarity, cold-start, … (dont reinvent the wheel)
  12. 12. acm recsys 2012:5 open problems
  13. 13. problem 1: predictionstemporality, multiple co-occurring objectives: diversity, novelty, freshness, serendipity, explainability
  14. 14. problem 2: algorithmsmore algorithms vs. more data vs. more rating effort
  15. 15. what is your algorithm doing? f(user, item) → R f(user, item1, item2) → R f(user, [item1...itemn]) → R e.g., @alexk_z @abellogin
  16. 16. problem 3: users + ratingssignals, context, groups, intents, interfaces
  17. 17. @xamat
  18. 18. problem 4: itemslifestyle, behaviours, decisions, processes, software development
  19. 19. @presnick
  20. 20. problem 5: measurementranking metrics vs. usability testing vs. A/B testing
  21. 21. Online Controlled Experiments: Introduction, Learnings, and HumblingStatisticshttp://www.exp-platform.com/Pages/2012RecSys.aspx
  22. 22. 3 key lessons
  23. 23. lesson 1: #recsys is an ensemble ...of disciplines statistics, machine learning, human-computer interaction, social network analysis, psychology
  24. 24. lesson 2: return to the domainuser effort, generative models, cost of a freakommendation, value you seek to create
  25. 25. @plamere
  26. 26. lesson 3: join the #recsys community learn, build, research, deploy: @MyMediaLite, @LensKitRS @zenogantner, @elehack contribute, read: #recsyswiki, @alansaid
  27. 27. recommender systems, today@neal_lathia

×