A Production-Ready RL Framework for Personalized Utility Tuning with Pareto Sweeping in Pinterest Recommender Systems

arXiv 2605.16344•2499a02752adb134d5adcc04bd4b890e6f181bcc60f328952b16eb53227a844c
LLM-adsad-auctionsbenchmarksbioinformaticscode-retrievaldeploymentevaluationfairnessindustrial-mljob-searchmulti-label-classificationpareto-frontierpersonalizationpopularity-biasproduction-systemsprotein-qarecommender-systemsreinforcement-learningretrieval-augmented-generationsoftware-engineering

Paper metadata

arXiv ID
2605.16344
Version
Not specified by this published record
Category
Computer Science — Information Retrieval (cs.IR)

The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.

Evidence and limitations

Source ID
arxiv_cs_ir
Record identifier
2499a02752adb134d5adcc04bd4b890e6f181bcc60f328952b16eb53227a844c
Enrichment time
2026-05-19T08:52:15Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.