Improving Conversational Passage Re-ranking with View Ensemble (SIGIR 23 Short Paper)

• Venue: SIGIR 2023 (short paper)
• Source: https://dl.acm.org/doi/abs/10.1145/3539618.3592002
• Topic area: Conversational Search

Key points

• The paper proposes a view ensemble strategy for conversational passage re-ranking, combining multiple views of the conversation context rather than relying on a single query reformulation or context representation.
• It targets the core conversational search challenge: user intent is spread across multiple dialogue turns, and fixed reformulations can lose information or introduce noise.
• The work fits into the modern search stack of retrieval → re-ranking → generation, where re-ranking quality directly affects downstream answer generation.

Method outline

As summarized in the forum post, methods in this problem space typically follow four stages:

1. Input and representation — encode queries, documents, and user context into dense or sparse representations, or structured prompts; 2. Core modules — retrievers, re-rankers, planners, memory modules, and tool interfaces, chained or run in parallel; 3. Learning strategy — supervised fine-tuning, contrastive learning, distillation, reinforcement learning, or synthetic data bootstrapping; 4. Inference strategy — single-pass retrieval, iterative retrieval, parallel sub-queries, early stopping, and budget control.

For this paper specifically, the ensembling of multiple views of the conversational context serves as a robustness mechanism at the re-ranking stage.

Evaluation context

• Datasets commonly used: MS MARCO, BEIR, Natural Questions, and domain-specific corpora
• Metrics: nDCG@10, MRR, Recall@k, Hit@k, human preference, task success rate, latency and token cost
• Baselines: BM25, dense retrieval, cross-encoder re-ranking, retrieval-free LLMs, commercial search APIs

> Note: Exact numerical results should be verified against the original PDF; this post is based on the abstract and public metadata.

Insights for search and recommendation systems

1. Architecture: cascade retrieval + re-ranking + generation remains mainstream, but agentic paradigms are making retrieval strategy itself learnable. 2. Data: high-quality instruction data and click/session logs matter as much as model architecture; 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. Productization: latency, cost, explainability, and safety are hard deployment constraints — do not optimize academic benchmarks alone.

Limitations and future work

Possible limitations include experiment scale limited by compute budget, benchmarks that diverge from real user distributions, unknown cross-lingual generalization due to English-centric data, and safety risks of agentic systems on the open web. Future directions include more efficient test-time compute allocation, deeper integration with knowledge graphs and structured databases, and causal/fairness constraints for recommendation.

Related work

• A Survey of Conversational Search, Sep 2025, ACM
• Engineering Conversational Search Systems: A Review of Applications (arXiv:2407.00997)
• ChatRetriever: Adapting Large Language Models for Generalized and Robust Retrieval (arXiv:2404.13556)
• CoSearchAgent: A Lightweight Collaborative Search Agent (arXiv:2402.06360)

Glossary

| Term | Meaning | |------|---------| | IR | Information Retrieval | | RAG | Retrieval-Augmented Generation | | LTR | Learning to Rank | | nDCG | Normalized Discounted Cumulative Gain | | Agentic Search | Modeling search as sequential decision-making and tool use | | Gen-IR | Generative Information Retrieval |