MaRI: Accelerating Ranking Model Inference via Structural Re-parameterization in Large Scale Recommendation System

TL;DR

MaRI is motivated by the observation that user-side computation is redundant in feature fusion matrix multiplication, and it therefore adopt the philosophy of structural reparameterization to alleviate such redundancy.

Abstract

Ranking models, i.e., coarse-ranking and fine-ranking models, serve as core components in large-scale recommendation systems, responsible for scoring massive item candidates based on user preferences. To meet the stringent latency requirements of online serving, structural lightweighting or knowledge distillation techniques are commonly employed for ranking model acceleration. However, these approaches typically lead to a non-negligible drop in accuracy. Notably, the angle of lossless acceleration by optimizing feature fusion matrix multiplication, particularly through structural reparameterization, remains underexplored. In this paper, we propose MaRI, a novel Matrix Re-parameterized Inference framework, which serves as a complementary approach to existing techniques while accelerating ranking model inference without any accuracy loss. MaRI is motivated by the observation that user-side computation is redundant in feature fusion matrix multiplication, and we therefore adopt the philosophy of structural reparameterization to alleviate such redundancy.

Figures (3)

• Figure 1: Inference Workflow of Our Online Large-Scale Recommendation System.
• Figure 2: Offline simulation results: Performance comparison (latency/FLOPs) between $\text{MatMul}_{\text{MaRI}}$ and vanilla MatMul under varying $B$, $D_{\text{user}}, D_{\text{item/cross}}$, and $D_{\text{hidden}}$).
• Figure 3: Fragmented MaRI vs. Vanilla MatMul and Neat MaRI.