Computing recommendations at extreme scale with Apache Flink @Buzzwords 2015

1. Till Rohrmann Flink committer / data Artisans trohrmann@apache.org @stsffap Computing Recommendations at Extreme Scale with Apache Flink

2. Recommendations: Collaborative Filtering 1

3. Recommendations §  Omnipresent nowadays §  Important for user experience and sales 2

4. Recommending Websites 3

5. Collaborative Filtering §  Recommend items based on users with similar preferences §  Latent factor models capture underlying characteristics of items and preferences of user §  Predicted preference: 4 ˆru,i = xu T yi

6. Rating Matrix §  Explicit or implicit ratings §  Prediction goal: Rating for unseen items 5 Items Users 10 5 ? ? 2 10 7 ? 5 ! " # # # $ % & & & Princess Facebook

7. Matrix Factorization §  Calculate low rank approximation to obtain latent factors 6 minX,Y ru,i − xu T yi( ) 2 + λ nu xu 2 + ni yi 2 i ∑ u ∑ # $ % & ' ( ru,i≠0 ∑ Items Users 10 5 ? ? 2 10 7 ? 5 ! " # # # $ % & & & ≈ X ! " # # # $ % & & & • Y( ) Princess Facebook §  = rating of user u for item i §  = latent factors of user u §  = latent factors of item i §  = regularization constant §  = number of rated items for user u §  = number of ratings for item i xu yi λ nu ni ru,i

8. Alternating Least Squares §  Hard to optimize since we have two variables §  Fixing one variables gives quadratic problem 7 X,Y xu = YSu YT + λnuΙ( ) −1 Yru T Sii u = 1 if ru,i ≠ 0 0 else " # $ %$ We only need the item vectors rated by user u

9. ALS Algorithm §  Update user matrix 8 Items Users 10 5 ? ? 2 10 7 ? 5 ! " # # # $ % & & & ≈ X ! " # # # $ % & & & • Y( ) Keep ﬁxed Calculate update

10. ALS Algorithm contd. §  Update item matrix 9 Items Users 10 5 ? ? 2 10 7 ? 5 ! " # # # $ % & & & ≈ X ! " # # # $ % & & & • Y( ) Keep ﬁxed Calculate update

11. ALS Algorithm contd. §  Repeat update step until convergence 10 Items Users 10 5 ? ? 2 10 7 ? 5 ! " # # # $ % & & & ≈ X ! " # # # $ % & & & • Y( ) Keep ﬁxed Calculate update

12. Apache Flink 11

13. What is Apache Flink? Apache Flink deep-‐dive by Stephan Ewen Tomorrow, 12:20 – 13:00 on Stage 3 12 Gelly Table FlinkML SAMOA DataSet (Java/Scala/Python) DataStream (Java/Scala) HadoopM/R Local Remote Yarn Tez Embedded Dataflow Dataflow(WiP) MRQL Table Cascading(WiP) Streaming dataflow runtime

14. Why Using Flink for ALS? §  Expressive API §  Pipelined stream processor §  Closed loop iterations §  Operations on managed memory 13

15. Expressive APIs §  DataSet: Abstraction for distributed data §  Computation speciﬁed as sequence of lazily evaluated transformations 14 case class Word(word: String, frequency: Int) val lines: DataSet[String] = env.readTextFile(…) lines.flatMap(line => line.split(“ “).map(word => Word(word, 1)) .groupBy(“word”).sum(“frequency”) .print()

16. Program Execution 15 case class Path (from: Long, to: Long) val tc = edges.iterate(10) { paths: DataSet[Path] => val next = paths .join(edges) .where("to") .equalTo("from") { (path, edge) => Path(path.from, edge.to) } .union(paths) .distinct() next } Optimizer Type extraction stack Task scheduling Dataflow metadata Pre-flight (Client) Master Workers Data Source orders.tbl Filter Map DataSourc e lineitem.tbl Join Hybrid Hash build HT prob e hash-part [0] hash-part [0] GroupRed sort forward Program Dataflow Graph deploy operators track intermediate results

17. Pipelined Stream Processor 16 Avoiding materializaBon of intermediate results

18. Iterate by looping §  for/while loop in client submits one job per iteration step §  Data reuse by caching in memory and/or disk Step Step Step Step Step Client 17

19. Iterate in the Dataﬂow 18

20. Memory Management 19

21. Memory Management 20

22. ALS implementations with Apache Flink 21

23. Naïve Implementation 1.  Join item vectors with ratings 2.  Group on user ID 3.  Compute new user vectors 22 xu = YSu YT + λnuΙ( ) −1 Yru T

24. Pros and Cons of Naïve ALS §  Pros •  Easy to implement §  Cons •  Item vectors are sent redundantly to network nodes •  Two shufﬂe steps make execution expensive 23

25. Blocked ALS Implementation 1.  Create user and item rating blocks 2.  Cache them on worker nodes 3.  Send all item vectors needed by user rating block bundled 4.  Compute block of item vectors 24 Based on Spark’s MLlib implementaBon

26. Pros and Cons of Blocked ALS §  Pros •  Reduces network load by avoiding data duplication •  Caching ratings: Only one shufﬂe step needed §  Cons •  Duplicates the rating matrix (user block/item block partitioning) 25

27. Performance Comparison 26 •  40 node GCE cluster, highmem-‐8 •  10 ALS iteraBons with 50 latent factors •  RaBng matrix has 28 billion non zero entries: Scale of NeAlix or SpoCfy

28. Machine Learning with FlinkML §  FlinkML contains blocked ALS §  Support for many other tasks •  Clustering •  Regression •  Classiﬁcation §  scikit-learn like pipeline support 27 val als = ALS() val ratingDS = env.readCsvFile[(Int, Int, Double)](ratingData) val parameters = ParameterMap() .add(ALS.Iterations, 10) .add(ALS.NumFactors, 50) .add(ALS.Lambda, 1.5) als.fit(ratingDS, parameters) val testingDS = env.readCsvFile[(Int, Int)](testingData) val predictions = als.predict(testingDS)

29. Closing 28

30. What Have You Seen? §  How to use collaborative ﬁltering to make recommendations §  Apache Flink, a powerful parallel stream processing engine §  How to use Apache Flink and alternating least squares to factorize really large matrices 29

31. 30

32. ﬂink.apache.org @ApacheFlink