AI & ML
From Relevance to Utility: Rubric Alignment in Yuanbao Search
arXiv cs.IR0 views
arXiv:2611.22503v1 Announce Type: new Abstract: Agentic search demands that a search engine move beyond per document relevance ranking to assemble document sets optimized for downstream LLM reasoning. Prior work addresses this either by improving individual pipeline stages in isolation or by applying RL based selection on top of relevance trained upstream modules, leaving the full stack unaligned to utility for reasoning. We propose rubric aligned search, a framework that unifies the traditional search pipeline under a single supervision surface. For each query, we derive a rubric from coarse to fine, specifying weighted positive criteria that the document set must satisfy and negative criteria that it must avoid. From this rubric we derive two training signals. First, a lightweight labeler distills rubric grounded judgments into per document utility labels, which supervise a contrastive dual encoder for retrieval and a multi task ranker trained with four rubric aligned scoring heads. Second, a rubric conditioned evaluator produces a set level reward that drives a dual branch RL selector, combining GRPO with best rollout distillation to resolve credit assignment in subset optimization. Deployed in Yuanbao Search, the pipeline improves end to end answer quality by 8.52% over the online production baseline.
Source:https://arxiv.org/abs/2611.22503PreviousMore Than a Score: Reinforcement Learning with Verbalized Feedback for Generative RecommendationNextMIPR: A Benchmark for Reasoning-Intensive Multimodal Intent-to-Product Retrieval with Multi-Category Solutions
Related posts
How OpenAI scales PostgreSQL to 800 million users
0dev.to Database
Spec-Driven PBT: Deriving Semantic Invariants from Code Specifications for Agentic Property-based Testing
0arXiv cs.SE
A Grey Literature Review of Graduate Students' Research Practices and Skills in the LLM Era
0arXiv cs.SE
ReBoT-MiCS: Reliability-Aware Bounding Time for Enhancing Timing Behavior of Embedded Mixed-Criticality Systems
0arXiv cs.SE
VoxelSBFL: Voxel-Level Spectrum-Based Failure Prediction for Vision-Language-Action Models
0arXiv cs.SE
Discontinuity-Aware Solvers for Vapor-Compression Cycles
0arXiv cs.NA