1: \begin{abstract}

2: Collaborative filtering algorithms are important building blocks in many practical recommendation systems.

3: For example, many large-scale data processing environments include collaborative filtering models for which the Alternating Least Squares (ALS) algorithm is used to compute latent factor matrix decompositions.

4: In this paper, we propose an approach to accelerate the convergence of parallel ALS-based optimization methods for collaborative filtering using a nonlinear conjugate gradient (NCG) wrapper around the ALS iterations. 

5: We also provide a parallel implementation of the accelerated

6: ALS-NCG algorithm in the Apache Spark distributed data

7: processing environment, and an efficient line search

8: technique as part of the ALS-NCG implementation that requires only one pass over the data on distributed datasets.

9: In serial numerical experiments on a linux workstation and

10: parallel numerical experiments on a 16 node cluster with 256

11: computing cores, we demonstrate that the combined ALS-NCG

12: method requires many fewer iterations and less time than

13: standalone ALS to reach movie rankings with high accuracy on 

14: the MovieLens 20M dataset. 

15: In parallel, ALS-NCG can achieve an acceleration factor of 4

16: or greater in clock time when an accurate solution is

17: desired; furthermore, the acceleration factor increases

18: as greater numerical precision is required in the

19: solution.

20: In addition, the NCG acceleration mechanism is efficient

21: in parallel and scales linearly with problem size on

22: synthetic datasets with up to nearly 1 billion ratings.

23: The acceleration mechanism is general and may also be

24: applicable to other optimization methods for collaborative

25: filtering.

26: \end{abstract}

27: