Metraipolis › archive › dc_9d136dee5c62

Embedding Subspace Partitioning for Dynamic Multi-Objective Retrieval — arXiv:2609.30601

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.30601
License (as archived)
CC0-1.0
Archived
2026-10-08T03:12:01.723Z
SHA-256 of the archived text
ae266f9c550e53bbbb9864fce51627a3a751f4e6a32c8159e4d18bffa5a59fe2
Stable URL of this page
https://metraipolis.com/documento/dc_9d136dee5c62

How to cite

Embedding Subspace Partitioning for Dynamic Multi-Objective Retrieval — arXiv:2609.30601 — arXiv (en), https://arxiv.org/abs/2609.30601 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9d136dee5c62 (Metraipolis archive dc_9d136dee5c62, sha256 ae266f9c550e53bbbb9864fce51627a3a751f4e6a32c8159e4d18bffa5a59fe2)
BibTeX
@misc{dc9d136dee5c62,
  title = {Embedding Subspace Partitioning for Dynamic Multi-Objective Retrieval — arXiv:2609.30601},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.30601},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9d136dee5c62; sha256 ae266f9c550e53bbbb9864fce51627a3a751f4e6a32c8159e4d18bffa5a59fe2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30601. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Modern industrial recommender systems must optimize across competing objectives, balancing semantic relevance with business metrics such as engagement and revenue. While bi-encoders dominate large-scale retrieval due to their efficiency, they collapse these heterogeneous signals into a single static embedding space. This design creates a fundamental limitation: once trained, the retriever cannot adapt to shifting objective priorities at serving time without retraining. Moreover, joint optimization with multi-objective losses often induces interference between objectives, leading to suboptimal trade-offs. We propose Embedding Subspace Partitioning (ESP), a retrieval framework that decomposes the embedding into task-aware subspaces and replaces the single dot product with a weighted sum of per-subspace similarities, whose weights are tunable at serving time. For Transformer bi-encoders, ESP uses the model's native end-of-sequence token as a segment delimiter, with segment-aware attention masking and position encoding resets to guarantee subspace isolation in a single forward pass. Serving is performed via GPU-accelerated exhaustive kNN over one concatenated index, eliminating the need for per-objective Approximate Nearest Neighbor (ANN) infrastructure required by multi-head approaches. We evaluate ESP on an open-source benchmark built from MS MARCO. A single ESP model traces a broad Pareto frontier, consistently outperforming strong multi-task baselines across diverse operating points. In LinkedIn's job matching platform (70M+ weekly users), ESP enabled dynamic retrieval reconfiguration and delivered significant key business metric lifts.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-28 in cs.IR).
Categorie arXiv: Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2609.30601 — 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 2026-10-08T02:50:01.725Z dalla pagina https://arxiv.org/catchup/cs.IR/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).