IAAILoopGR
LoopGR: Parameter-Efficient Scaling for Generative Recommendation via Recurrent Depth
Brings recurrent depth to generative recommendation with a hierarchical mechanism that alternates a shared generative decoder and a lightweight context updater, plus recurrent trajectory self-distillation from a high-budget teacher to a low-budget student.
Inside the method
[ Shared generative decoder ↔ context updater ] × r → next item
Simplified conceptual schematic. Consult the paper for the complete architecture.
- Recurrence family
- Core recurrence
- Depth control
- Configurable recurrent budget; distilled low-budget student
- KV / state strategy
- See paper
Reading note
Outperforms strong generative recommendation baselines and beats an encoder-decoder baseline while reducing model size by up to 52%.
Sources checked 2026-10-28. This catalog does not imply independent reproduction.
Related works
Cite this work
@misc{loopgr2026,
title = {LoopGR: Parameter-Efficient Scaling for Generative Recommendation via Recurrent Depth},
author = {Huizhe Zhang and Yudong Li and Liang Chen and Chengxiang Zhuo and Zang Li},
year = {2026},
eprint = {2610.34157},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.34157}
}