Lune

ICML2025Top-tier venue

Retrieval Augmented Time Series Forecasting

Sungwon Han, Seungeon Lee, Meeyoung Cha, Sercan Ö. Arik, Jinsung Yoon

Abstract

Time series forecasting uses historical data to predict future trends, leveraging the relationships between past observations and available features. In this paper, we propose RAFT, a retrieval-augmented time series forecasting method to provide sufficient inductive biases and complement the model's learning capacity. When forecasting the subsequent time frames, we directly retrieve historical data candidates from the training dataset with patterns most similar to the input, and utilize the future values of these candidates alongside the inputs to obtain predictions. This simple approach augments the model's capacity by externally providing information about past patterns via retrieval modules. Our empirical evaluations on ten benchmark datasets show that RAFT consistently outperforms contemporary baselines with an average win ratio of 86%. Over the past decade, deep learning models such as CNNs (Bai et al., 2018; Borovykh et al., 2017) and RNNs (Hewamalage et al., 2021) have proven their effectiveness in capturing patterns of change in historical

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

17 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Improving Language Models by Retrieving from Trillions of Tokens cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
  • Improving Language Models by Retrieving from Trillions of Tokens cites Retrieval Augmented Language Model Pre-Training
  • TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
  • Enhancing Multivariate Time Series Forecasting with Global Temporal Retrieval cites TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster
  • Enhancing Multivariate Time Series Forecasting with Global Temporal Retrieval cites Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting
  • Enhancing Multivariate Time Series Forecasting with Global Temporal Retrieval cites Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution Shift
  • Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting cites Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution Shift
  • Improving Time Series Forecasting via Instance-aware Post-hoc Revision cites Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting
  • Improving Time Series Forecasting via Instance-aware Post-hoc Revision cites Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution Shift
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers17

  1. TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot ForecasterKanghui Ning, Zijie Pan, Yu Liu, Yushan Jiang et al.NeurIPS 2025 · 53 citations
  2. Improving Time Series Forecasting via Instance-aware Post-hoc RevisionZhiding Liu, Mingyue Cheng, Guanhao Zhao, Jiqian Yang et al.NeurIPS 2025 · 16 citations
  3. STRAP: Spatio-Temporal Pattern Retrieval for Out-of-Distribution GeneralizationHaoyu Zhang, Wentao Zhang, Hao Miao, Xinke Jiang et al.NeurIPS 2025 · 12 citations
  4. MemCast: Memory-Driven Time Series Forecasting with Experience-Conditioned ReasoningXiaoyu Tao, Mingyue Cheng, Ze Guo, Shuo Yu et al.ICML 2026 · 8 citations
  5. Enhancing Multivariate Time Series Forecasting with Global Temporal RetrievalFanpu Cao, Lu Dai, Jindong Han, Hui XiongICLR 2026 · 7 citations

Builds on13

  1. Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni et al.NeurIPS 2020 · 19,162 citations
  2. Retrieval Augmented Language Model Pre-TrainingKelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat et al.ICML 2020 · 2,937 citations
  3. Improving Language Models by Retrieving from Trillions of TokensSebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai et al.ICML 2022 · 1,629 citations
  4. Non-stationary Transformers: Exploring the Stationarity in Time Series ForecastingYong Liu, Haixu Wu, Jianmin Wang, Mingsheng LongNeurIPS 2022 · 1,080 citations
  5. Reversible Instance Normalization for Accurate Time-Series Forecasting against Distribution ShiftTaesung Kim, Jinhee Kim, Yunwon Tae, Cheonbok Park et al.ICLR 2022 · 1,020 citations

Related papers

  1. Stationarity-Aware Retrieval-Augmented Time Series ForecastingShiqiao Zhou, Holger Schöner, Zipeng Wu, Edouard Fouché et al.KDD 2026 · 1 citation
  2. Predicting the Future by Retrieving the PastDazhao Du, Tao Han, Song GuoAAAI 2026
  3. QuiZSF: A Retrieval-Augmented Framework for Zero-Shot Time Series ForecastingShichao Ma, Zhengyang Zhou, Qihe Huang, Binwu Wang et al.WWW 2026
  4. Retrieval-Augmented Diffusion Models for Time Series ForecastingJingwei Liu, Ling Yang, Hongyan Li, Shenda HongNeurIPS 2024 · 62 citations
  5. TimeMRA: LLM-Empowered Time Series Forecasting via Multi-Scale Retrieval-Augmented RepresentationsZongjiang Shang, Chengxi Jin, Binqing Wu, Dongliang Cui et al.ICML 2026
Dusk over the sea between two cliffs drawn in fine vertical lines