MétaCan
Menu
Back to cohort
Record W4297855566 · doi:10.48550/arxiv.1508.03110

Algorithmic Acceleration of Parallel ALS for Collaborative Filtering:\n Speeding up Distributed Big Data Recommendation in Spark

2015· preprint· en· W4297855566 on OpenAlexaff
Manda Winlaw, Michael B Hynes, Anthony L. Caterini, Hans De Sterck

Bibliographic record

VenuearXiv (Cornell University) · 2015
Typepreprint
Languageen
FieldComputer Science
TopicRecommender Systems and Techniques
Canadian institutionsUniversity of Waterloo
Fundersnot available
KeywordsMovieLensComputer scienceCollaborative filteringSPARK (programming language)AccelerationConjugate gradient methodParallel computingWorkstationSpeedupLine searchComputational scienceAlgorithmRecommender systemMachine learningPath (computing)

Abstract

fetched live from OpenAlex

Collaborative filtering algorithms are important building blocks in many\npractical recommendation systems. For example, many large-scale data processing\nenvironments include collaborative filtering models for which the Alternating\nLeast Squares (ALS) algorithm is used to compute latent factor matrix\ndecompositions. In this paper, we propose an approach to accelerate the\nconvergence of parallel ALS-based optimization methods for collaborative\nfiltering using a nonlinear conjugate gradient (NCG) wrapper around the ALS\niterations. We also provide a parallel implementation of the accelerated\nALS-NCG algorithm in the Apache Spark distributed data processing environment,\nand an efficient line search technique as part of the ALS-NCG implementation\nthat requires only one pass over the data on distributed datasets. In serial\nnumerical experiments on a linux workstation and parallel numerical experiments\non a 16 node cluster with 256 computing cores, we demonstrate that the combined\nALS-NCG method requires many fewer iterations and less time than standalone ALS\nto reach movie rankings with high accuracy on the MovieLens 20M dataset. In\nparallel, ALS-NCG can achieve an acceleration factor of 4 or greater in clock\ntime when an accurate solution is desired; furthermore, the acceleration factor\nincreases as greater numerical precision is required in the solution. In\naddition, the NCG acceleration mechanism is efficient in parallel and scales\nlinearly with problem size on synthetic datasets with up to nearly 1 billion\nratings. The acceleration mechanism is general and may also be applicable to\nother optimization methods for collaborative filtering.\n

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow)
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Methods · Consensus signal: none
Teacher disagreement score0.962
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.001
Science and technology studies0.0000.000
Scholarly communication0.0000.001
Open science0.0020.002
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.403
GPT teacher head0.276
Teacher spread0.128 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one teacher head, not a consensus.

Study designSimulation or modeling
Domainnot available
GenreMethods

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2015
Admission routes1
Has abstractyes

Explore more

Same venuearXiv (Cornell University)Same topicRecommender Systems and TechniquesFrench-language works237,207