Aligned Query Expansion: Efficient Query Expansion for Information Retrieval through LLM Alignment (arXiv, July 2025)
Paper Information
| Field | Content | |------|------| | Title | Aligned Query Expansion: Efficient Query Expansion for Information Retrieval through LLM Alignment | | Authors | Adam Yang, Gustavo Penha, Enrico Palumbo, Hugues Bouchard | | Published | July 2025, arXiv | | Link | https://arxiv.org/abs/2507.11042 | | Category | Query Understanding |
Overview
The paper tackles query expansion in information retrieval. Large language models can improve expansion quality and query intent understanding, but naive LLM-based expansion adds inference latency and token cost at query time. This work proposes aligning an LLM for query expansion so the model becomes more efficient while retaining retrieval effectiveness, targeting scenarios such as open-domain search, enterprise knowledge retrieval, and conversational search.
Position in the IR Landscape
The post places the paper within the evolution of neural information retrieval:
- Classic stack: BM25 recall → cross-encoder/dense reranking → generation, a cascade balancing coverage, discrimination, and presentation.
- Dense retrieval: millisecond-level recall via approximate nearest-neighbor search, but sensitive to domain shift and long-tail queries.
- Cross-encoders: high precision but no pre-computable document representations.
- Generative and agentic retrieval: iterative, verifiable, plannable retrieval, but with index-update and inference-cost challenges.
- Datasets: MS MARCO, BEIR, Natural Questions, domain-specific corpora
- Metrics: nDCG@10, MRR, Recall@k, Hit@k, latency and token cost
- Baselines: BM25, dense retrieval, cross-encoder reranking, retrieval-free LLMs
- Beyond the limitation of a single query: Train your LLM for query expansion (arXiv:2510.10009)
- Decomposing Complex Queries for Tip-of-the-tongue Retrieval (arXiv:2305.15053)
- Few-Shot Generative Conversational Query Rewriting (SIGIR 2020)
- Hypothetical Documents or Knowledge Leakage? Rethinking LLM-based Query Expansion (arXiv:2504.14175)
- LLM-Based Query Expansion with Gaussian Kernel Semantic Enhancement
The key new variables in the LLM era are the inference budget and the action space (whether to retrieve, how many times, and which tools to call). Aligned query expansion reduces the cost of the expansion step within this budget.
Typical Experimental Setup (for context)
Specific numeric results should be verified against the original PDF; this post is based on the abstract and public metadata.
Engineering Checklist
| Item | Question | Recommendation | |--------|----------|----------------| | Data | Does training/indexing contain PII? Version control? | Partitioned indexes, anonymization, rollback-capable embedding versions | | Latency | What is the p99 budget? How many retrieval steps? | Cascade + early stopping, hot-query caching, async reranking | | Quality | Do offline gains translate to online CTR/satisfaction? | Interleaving experiments, human audits, citation checks | | Safety | Does open retrieval introduce poisoning/bias? | Source whitelists, adversarial detection, output filtering | | Cost | Per-query token and GPU usage? | Small-model routing, distillation, hybrid sparse+dense |
Takeaways
1. Architecture: cascaded retrieval + reranking + generation remains mainstream, but agentic approaches make the retrieval policy itself learnable; efficient, aligned query expansion cuts the latency/cost of that policy. 2. Data: high-quality instruction data and click/session logs are as important as model design. 3. Evaluation: the gap between offline metrics and online satisfaction is widening; LLM-as-judge needs cross-validation with human assessment. 4. Deployment: latency, cost, interpretability, and safety are hard constraints — do not optimize academic benchmarks alone.
Related Entries
Glossary
| Term | Meaning | |------|---------| | IR | Information Retrieval | | RAG | Retrieval-Augmented Generation | | nDCG | Normalized Discounted Cumulative Gain, a ranking quality metric | | Agentic Search | Modeling search as sequential decision-making and tool calls | | Gen-IR | Generative Information Retrieval |
Original Abstract
> Aligned Query Expansion: Efficient Query Expansion for Information Retrieval through LLM Alignment — see arXiv:2507.11042.