COS-Mix: Cosine Similarity and Distance Fusion for Improved Information Retrieval
Paper: arXiv:2406.00638 — Kush Juvekar, Anupam Purwar (June 2024)
Overview
COS-Mix proposes fusing cosine similarity and Euclidean distance into a single combined metric for semantic/vector search. Rather than relying on one similarity measure alone, the fused score aims to better capture the geometric relationship between query and document embeddings, improving retrieval quality in dense retrieval pipelines.
Context: Hybrid Search vs. Vector Search
The forum discussion places this work in the broader landscape of modern information retrieval:
- Classic search stacks act as a funnel: recall for coverage, ranking for discrimination, and generation for presentation. In the LLM era, new variables include inference budget and action space (whether to retrieve, how many times, which tools to call).
- Neural IR has evolved from BM25 → BERT cross-encoders → bi-encoder dense retrieval → late interaction → generative retrieval and LLM agents, each balancing efficiency, effectiveness, and maintainability.
- Dense retrieval achieves millisecond-level recall via approximate nearest-neighbor search but is sensitive to domain shift and long-tail queries; cross-encoders are accurate but cannot precompute document representations.
- RAG and agentic search extend external knowledge access from one-shot retrieval to iterative, verifiable, plannable processes, shifting evaluation from static nDCG toward task success rate, citation accuracy, and multi-hop reasoning completeness.
- Datasets: MS MARCO, BEIR, Natural Questions, domain-specific corpora
- Metrics: nDCG@10, MRR, Recall@k, Hit@k, human preference, task success rate, latency, token cost
- Baselines: BM25, dense retrieval, cross-encoder reranking, retrieval-free LLMs, commercial search APIs
- Azure AI Search: Outperforming vector search with hybrid retrieval
- Deep Retrieval at CheckThat! 2025: Identifying Scientific Papers
- Domain-specific Question Answering with Hybrid Search (arXiv:2412.03736)
- Efficient Knowledge Graph Construction and Retrieval from Unstructured Data (arXiv:2507.03226)
- Hybrid Hierarchical Retrieval for Open-Domain Question Answering
- Modernizing Facebook Scoped Search: Keyword and Embedding Hybrid Retrieval (arXiv:2509.13603)
Evaluation Practices Discussed
The post notes that quantitative results should be verified against the original PDF before citation.
Engineering Checklist
| Concern | Question | Recommendation | |---|---|---| | Data | PII in training/index? Versioning? | Partitioned indexes, anonymization, rollback-capable embedding versions | | Latency | p99 budget? Retrieval steps? | Cascades + early stopping, hot-query caching, async reranking | | Quality | Do offline gains transfer online? | Interleaving experiments, human audits, citation checks | | Safety | Poisoning/bias in open retrieval? | Source whitelists, adversarial detection, output filtering | | Cost | Per-query token/GPU spend? | Small-model routing, distillation, hybrid sparse+dense |
Takeaways
1. Architecture: Cascaded retrieve–rerank–generate remains mainstream, but agentic paradigms are making retrieval policy itself learnable. 2. Data: High-quality instruction data and click/session logs matter; synthetic data risks leakage and distribution shift. 3. Evaluation: The gap between offline metrics and online satisfaction is widening; LLM-as-judge needs cross-validation with human evaluation. 4. Production: Latency, cost, explainability, and safety are hard constraints—do not optimize academic benchmarks alone.