Improving Conversational Passage Re-ranking with View Ensemble (SIGIR 23 Short Paper)
Metadata
| Field | Value | |------|------| | Title | 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 | | Resource type | Academic paper | | Category | Conversational Search |
Background and Motivation
Conversational search systems must re-rank retrieved passages under multi-turn user context, evolving intent, and strict latency constraints. Traditional pipelines that treat retrieval, ranking, and generation as disjoint stages struggle to meet the demands of natural-language interaction, multi-hop reasoning, and up-to-date knowledge in the LLM era. This SIGIR 2023 short paper targets the conversational passage re-ranking stage, proposing a view ensemble approach that combines multiple views/representations of the conversation context to improve re-ranking effectiveness.
Positioning in the Field
The paper sits at the intersection of conversational search and large-scale search/recommendation systems. Viewed through the classic search funnel — recall (coverage), re-ranking (discrimination), generation (presentation) — the LLM era adds new variables: inference budget and action space (whether to retrieve, how many times, which tools to call).
Neural IR has evolved from BM25 to BERT cross-encoders, bi-encoder dense retrieval, late interaction, and now generative retrieval and LLM agents, each generation balancing efficiency–effectiveness–maintainability. Dense retrieval enables 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; generative approaches reduce cascade errors but face index-update challenges.
Relevant Context for Practitioners
- Architecture: cascaded retrieval + re-rank + generate remains mainstream, but the agentic paradigm treats retrieval count and strategy as learnable objects.
- Evaluation: offline metrics increasingly diverge from online satisfaction; LLM-as-judge should be cross-validated with human assessment.
- Engineering checklist: watch index update frequency, embedding version compatibility, A/B sensitivity, and failure modes (empty retrieval, wrong tool calls, over-generation).
- A Survey of Conversational Search (Sep 2025, ACM)
- Engineering Conversational Search Systems: A Review of Applications
- Agentic Conversational Search with Contextualized Reasoning via Reinforcement Learning
- CTR-Guided Generative Query Suggestion in Conversational Search (EMNLP)
- ChatRetriever: Adapting LLMs for Generalized and Robust Conversational Retrieval
- CoSearchAgent: A Lightweight Collaborative Search Agent with LLMs
- Original paper: *Improving Conversational Passage Re-ranking with View Ensemble* — ACM DL
Note: quantitative results should be verified against the original PDF; this entry is based on the abstract and public metadata.