Building and Deploying a Multi-Stage Recommender System with Merlin

Higley, Karl; Oldridge, Even; Ak, Ronay; Rabhi, Sara; Moreira, Gabriel de Souza Pereira

doi:10.1145/3523227.3551468

Cited by 6 publications

(1 citation statement)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When dealing with real-world problems, it is possible to identify a common architectural pattern used by companies when deploying large-scale multi-stage recommendation systems models which comprises of two major components: A recommender system model, capable of generating candidate items, followed by a re-ranker model, which generates the final list of recommendations for each user, as proposed to some extent in papers and blogs by NVIDIA (Higley et al 2022), Google (Covington, Adams, and Sargin 2016), Meta's Instagram (Medvedev, Wu, and Gordon…”

Section: Introductionmentioning

confidence: 99%

Addressing Fast Changing Fashion Trends in Multi-Stage Recommender Systems

Roy¹,

D’Amico²,

Lawlor³

et al. 2023

FLAIRS

View full text Add to dashboard Cite

Fashion industry is driven by fashion cycles, in which a fashion item is launched, rises to mainstream appeal and becomes a trend, then diminishes and eventually becomes obsolete. These properties make it critical to incorporate temporal information when adapting a recommendation framework to be employed in the fashion domain. However, an industry standard real-world recommendation architecture entails numerous phases, including data preparation, establishing and training recommender models, filtering and fulfilling revenue-based user needs. The contributions of the presented work are twofold. We first formalise the multi-stage recommendation pipeline by including the time dimension intrinsically present in the fashion data. We then present a study to incorporate explicit fashion domain characteristics into the presented pipeline. Finally, we conduct comprehensive experimentation on a real-world web-scale fashion dataset released by H\&M, illustrating how including domain knowledge in the multi-stage framework can lead to significantly improvement on the final recommendation performance.

show abstract

Section: Introductionmentioning

confidence: 99%