AdaTutoRank: Learning to Rerank Document Sets via Adaptive Tutoring Optimization for RAG and Deep Research
cs.CL
Submitted: 2026-09-26
Updated: 2026-09-29
Comments: Project Page: https://adatutorank.github.io/
Code: https://github.com/AdaTutoRank/AdaTutoRank
Project page: https://adatutorank.github.io
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
The gist: Document rerankers determine what evidence reaches the downstream model in RAG and deep research, yet mainstream rerankers select by relevance matching, and individually relevant documents rarely
Terminology
Abstract
Document rerankers determine what evidence reaches the downstream model in RAG and deep research, yet mainstream rerankers select by relevance matching, and individually relevant documents rarely constitute the complete, complementary, non-redundant set a complex information need demands. Prior work rewards a set by its aggregate rubric score, shifting the objective from ranking documents to composing sets. Yet that score is one scalar shared by every document in the set, so the supervision is sparse: a redundant document is rewarded with the rest whenever the set scores well, and a decisive one penalized with the rest whenever it does not; credit assignment leaves contributors indistinguishable from free riders. On-policy distillation could densify this supervision, but existing methods give every rollout the same fixed guidance, too prescriptive for strong rollouts and too abstract for weak ones. We therefore propose AdaTutoRank, a setwise reranker trained with Adaptive Tutoring Optimization (ATO) under a three-level hierarchy of nine rubric dimensions, which supplies silver labels for the cold start, rewards for reinforcement learning, and hints for distillation. ATO draws three hint forms of increasing specificity from the policy's own frozen snapshot: the rubrics alone, a self-selector's sibling-set chosen under rubrics, and a self-reflector's reflection contrasting the rollout with that sibling-set; each rollout receives the form matched to its quality. Re-scoring that rollout under the hint-conditioned frozen teacher and the hint-free snapshot distills the hint's effect into a token-level advantage that complements the group-relative outcome advantage. Across ten benchmarks spanning RAG, deep research, and setwise evaluation, AdaTutoRank attains the best overall performance while issuing fewer retrieval calls.
Sources
- HealthBench: Evaluating Large Language Models Towards Improved Human Health
- OpenScholar: Synthesizing Scientific Literature with Retrieval-augmented LMs
- Rank4Gen: RAG-Preference-Aligned Document Set Selection and Ranking
- Can Multimodal Large Language Models Understand OCT?
- Retrieval-Augmented Generation for Large Language Models: A Survey
- Reinforcement Learning via Self-Distillation
- Beyond Relevance-Centric Retrieval: Rubric-Oriented Document Set Selection and Ranking
- Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
- KnowHal: A Knowledge-Driven Benchmark for Comprehensive Multimodal Hallucination Evaluation
- From Proprietary to Open-Source: Bridging the Distribution Gap via Multi-Agent Protocol Distillation in Agentic Search
- Training Documents Reranker with Search Rubrics for Deep Research Agent
- Can Visual Input Be Compressed? A Visual Token Compression Benchmark for Large Multimodal Models
- RankVicuna: Zero-Shot Listwise Document Reranking with Open-Source Large Language Models
- RankZephyr: Effective and Robust Zero-Shot Listwise Reranking is a Breeze!
- DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Self-Distillation Enables Continual Learning
- Deep Research: A Systematic Survey
- A Comprehensive Survey of Deep Research: Systems, Methodologies, and Applications
- ReAct: Synergizing Reasoning and Acting in Language Models
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering