Open access
Aug 2026
SETRec++: Scaling Order-agnostic Identifier for Large Language Model-based Generative Recommendation
A novel set identifier paradigm is introduced, representing each item as a set of order-agnostic tokens, which proves SETRec’s superior efficiency and scalability on cold-start items as model sizes increase, and SETRec++’s potential in the scalability of order-agnostic identifier.
Xin-Yu Lin, Chuan-Bo Zhang, Yu-Fan Liu et al.
· ACM Transactions on Recommen... · 0 citations