May 27, 2021 · Artificial Intelligence

Explicit Semantic Cross Feature Learning via Pre-trained Graph Neural Networks for CTR Prediction (PCF‑GNN)

PCF‑GNN builds a heterogeneous graph of feature nodes and learns edge statistics via pre‑training, enabling it to infer unseen cross‑features, reduce storage by over 50%, and consistently improve CTR prediction AUC compared to implicit and explicit baselines, with proven online gains.

Graph Neural NetworkRecommendation Systemscross feature

0 likes · 12 min read

Explicit Semantic Cross Feature Learning via Pre-trained Graph Neural Networks for CTR Prediction (PCF‑GNN)