Metraipolis › archive › dc_653b0eae350c

Merge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for Recommendation — arXiv:2610.37973

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.37973
License (as archived)
CC0-1.0
Archived
2026-10-31T06:11:12.460Z
SHA-256 of the archived text
41265d147ffba2c9f3be043ad9c596611e51de3a92f6889055b8b1c0ca1ad74b
Stable URL of this page
https://metraipolis.com/documento/dc_653b0eae350c

How to cite

Merge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for Recommendation — arXiv:2610.37973 — arXiv (en), https://arxiv.org/abs/2610.37973 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_653b0eae350c (Metraipolis archive dc_653b0eae350c, sha256 41265d147ffba2c9f3be043ad9c596611e51de3a92f6889055b8b1c0ca1ad74b)
BibTeX
@misc{dc653b0eae350c,
  title = {Merge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for Recommendation — arXiv:2610.37973},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.37973},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_653b0eae350c; sha256 41265d147ffba2c9f3be043ad9c596611e51de3a92f6889055b8b1c0ca1ad74b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.37973. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM-based generative recommendation reads long, text-rich interaction histories but generates only short semantic identifiers (SIDs). This asymmetry makes supervised fine-tuning inefficient: encoding the history is expensive, yet each history usually provides supervision for only one target. We present Merge and Branch (M&B) to reduce input cost and learn more from each encoded history. Merge compresses each historical item's SID and text into a small number of learned tokens. Branch uses the compressed history to predict multiple target SIDs. To share history computation while keeping these targets independent, M&B organizes each example as a shallow, wide tree. We develop tree FlashAttention to execute this layout efficiently, processing the shared history once and propagating gradients from every branch for end-to-end training. A tree-structured key/value cache extends history sharing to constrained decoding. Experiments on public benchmarks show that M&B outperforms strong recommendation baselines while substantially improving training and decoding efficiency.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Information Retrieval (cs.IR) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.37973 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T06:07:08.611Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).