From Language to Behavior: Scaling Sequence Transformers for Industrial Recommendation Ranking with Rec-Native Designs
ReST is proposed, a recommendation-native Transformer scaling framework that achieves higher accuracy and scales more consistently along sequence length, depth, and width, where LLM-style Transformer blocks saturate.
Jie Chen, Xiang-Qia Yu, Yan-Chao Lian et al.
· 0 citations