Q-TIE: A Lightweight and Generalizable Re-ranking Framework for Temporal Information Retrieval
cs.CL, cs.IR
Submitted: 2026-09-20
Updated: 2026-09-20
Comments: Accepted to EMNLP 2026 (Main Conference). Code: https://github.com/ssoy0701/Q-TIE
Code: https://github.com/ssoy0701/Q-TIE
License: http://creativecommons.org/licenses/by/4.0/
The gist: Temporal Information Retrieval (TIR) has been increasingly critical given the rise of Retrieval-Augmented Generation (RAG).
Terminology
Abstract
Temporal Information Retrieval (TIR) has been increasingly critical given the rise of Retrieval-Augmented Generation (RAG). Since temporally mismatched evidence can be highly misleading, TIR aims to retrieve documents that are both semantically and temporally relevant to a query. Two TIR paradigms have emerged - temporal retrievers and temporal re-rankers - differing in how temporal relevance is modeled. While these paradigms provide complementary strengths, our analysis reveals that each alone falls short of robust TIR: temporal retrievers provide flexible query understanding via learned representations, but often fail to explicitly account for temporal constraints; temporal re-rankers can enforce such constraints more explicitly, but often rely on predefined re-ranking rules. To address this, we propose Q-TIE, a re-ranking framework based on learned Temporal Intent Extraction (TIE). By introducing a TIE model that maps each query's temporal constraint into a unified interval representation (i.e., t start, t end), Q-TIE generalizes beyond predefined rules via model-based learning while explicitly modeling temporal constraints as a separate signal - jointly achieving what each paradigm typically trades off. Experiments demonstrate that Q-TIE consistently outperforms existing TIR methods with stronger generalizability across temporal query types, and provides a lightweight yet effective add-on for temporally-aware RAG pipelines. Code: https://github.com/ssoy0701/Q-TIE.
Sources
- Retrieval-Augmented Generation for Large Language Models: A Survey
- Gemma 2: Improving Open Language Models at a Practical Size
- Gemma 3 Technical Report
- TempRetriever: Fusion-based Temporal Dense Passage Retrieval for Time-Sensitive Questions
- The Llama 3 Herd of Models
- Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- RAG Meets Temporal Graphs: Time-Sensitive Modeling and Retrieval for Evolving Knowledge
- A Dataset for Answering Time-Sensitive Questions
- Mistral 7B
- Nomic Embed: Training a Reproducible Long Context Text Embedder
- It's High Time: A Survey of Temporal Question Answering
- SQuAD: 100,000+ Questions for Machine Comprehension of Text
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- MRAG: A Modular Retrieval Framework for Time-Sensitive Question Answering
- Temporal Knowledge Graph Question Answering: A Survey
- Towards Benchmarking and Improving the Temporal Reasoning Capability of Large Language Models
- Corrective Retrieval Augmented Generation
- Qwen3 Technical Report
- MetaMath: Bootstrap Your Own Mathematical Questions for Large 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