Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

TimeRAF: Retrieval-Augmented Foundation model for Zero-shot Time Series Forecasting

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read TimeRAF retrieves similar time series from a knowledge base to improve zero-shot forecasting of a frozen foundation model.

desk verdict A clean retrieval-augmentation idea on a frozen TSFM, with small reported gains and reporting choices that make the magnitude hard to trust; worth refereeing if the authors tighten the evidence. read the letter →

arxiv 2412.20810 v1 pith:QQNBEOD6 submitted 2024-12-30 cs.LG

Huanyu Zhang , Chang Xu , Yi-Fan Zhang , Zhang Zhang , Liang Wang , Jiang Bian , Tieniu Tan This is my paper · ORCID
classification cs.LG
keywords timeseriesforecastingzero-shotlearningretrieval-augmentedgenerationfoundationmodelslearnableretrieverchannelpromptingknowledgebasefrozenbackbone
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

TimeRAF tries to establish that retrieval-augmented generation, already common in language and vision, can also make zero-shot time series forecasting better without retraining the underlying foundation model. The paper proposes a learnable retriever that selects the most useful series from a multi-domain knowledge base, and Channel Prompting, which injects their embeddings into a frozen TTM-Base backbone. In experiments across six standard benchmarks (ETTh1, ETTh2, ETTm1, ETTm2, Weather, Electricity), the retrieval-augmented model reduces zero-shot mean squared error on every dataset compared with the frozen backbone alone. The point of the paper is that external knowledge, retrieved on the fly, is a cheap complement to what a pre-trained forecaster already knows.

What carries the argument

The load-bearing mechanism is the pair formed by a dual-encoder retriever and Channel Prompting. The retriever embeds the input query and each candidate window from a knowledge base of roughly three million 512-step series (drawn from a 320-million-point subset of the LOTSA and UTSD corpora), scores candidates by dot-product similarity, and keeps the top k; it is trained by minimizing the KL divergence between its scores and a distribution built from how much each candidate improves the frozen forecaster's prediction error. Channel Prompting then flattens and concatenates the input embedding with each retrieved candidate embedding, compresses the concatenation through an MLP, averages the k refinements, and adds the result back to the input embedding as a residual before the backbone forecasts. This lets the foundation model consume external knowledge without any change to its own weights.

What would settle it

Remove every knowledge-base window whose timestamps overlap the test split or the 96-step forecast horizon of any of the six evaluation datasets, rerun TimeRAF's zero-shot evaluation, and compare MSE to the frozen backbone; if the gains collapse or vanish, the reported improvement comes from retrieving test-domain data rather than from transferable external knowledge.

Watch

Extended reading notes

Core claim

The paper's central claim is that a frozen time-series foundation model can be upgraded for zero-shot forecasting by retrieving relevant series from an external knowledge base and merging them into the input representation. With TTM-Base kept frozen, TimeRAF reports lower MSE than the backbone on all six test sets (for example ETTh1 drops from 0.364 to 0.359, Weather from 0.158 to 0.152, Electricity from 0.170 to 0.168), and it also edges out several existing zero-shot and full-shot baselines. The ablations attribute the gain to both components: a retriever trained end-to-end to prefer candidates that actually improve the forecaster's metric beats random selection and cosine-similarity selection, and Channel Prompting beats token-level concatenation and plain embedding averaging as the integration method.

Load-bearing premise

The load-bearing premise is that the six evaluation datasets are not represented in the knowledge base; the paper never states that ETTh1, ETTh2, ETTm1, ETTm2, Weather, or Electricity were excluded from the ~3-million-point subset drawn from LOTSA and UTSD, the same large public corpora used to build the training data.

Editorial extensions

If this is right

  • Zero-shot forecasting can be improved without retraining or fine-tuning the foundation model, so an already-deployed forecaster can be upgraded just by adding a retriever, a knowledge base, and the prompting module.
  • Retrieval quality should be judged by the downstream forecast metric rather than by raw similarity; the paper's retriever learns to rank candidates by usefulness, and cosine-similarity retrieval is consistently worse.
  • The integration method matters: prompt injections that compress concatenated input and candidate embeddings along the channel dimension outperform token-level concatenation and simple averaging in the paper's ablations.
  • The knowledge base can be adapted at inference time, with a curated multi-domain base giving the best results overall and a smaller domain-matched base becoming competitive once the multi-domain base is shrunk.
  • As a plug-in on top of a backbone, TimeRAF inherits future improvements to the backbone itself; the paper states that gains should grow as the foundation model evolves.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct test the paper leaves implicit is full domain hold-out: build the knowledge base from every domain except healthcare, then evaluate on held-out healthcare series; if the gains persist, the retriever is transferring structure across domains rather than memorizing near-duplicates.
  • The retriever and Channel Prompting module are trained only against TTM-Base, so an open question is whether the same retrieval scores and prompting weights transfer to a different frozen backbone; the paper's own expectation that TimeRAF improves with better backbones implies this should be tested.
  • Because the knowledge base is built with sliding windows over corpora that include the classic evaluation datasets, a fair audit should remove any knowledge-base window whose timestamps fall inside the test horizons of the six benchmarks; the paper does not report such an exclusion check.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This paper proposes TimeRAF, a retrieval-augmented framework for zero-shot multivariate time-series forecasting. A frozen TSFM backbone (TTM-Base) is augmented by (i) a dual-encoder MLP retriever that scores 512-length windows in a multi-domain knowledge base and selects the top-k candidates, (ii) a Channel Prompting MLP that integrates the retrieved candidate embeddings into the input embedding, and (iii) an end-to-end training loss combining a prediction loss with a KL-divergence retrieval loss. The knowledge base is built by sliding-window preprocessing of roughly 3M points selected from LOTSA/UTSD training subsets. On six standard benchmarks (ETTh1/ETTh2/ETTm1/ETTm2, Weather, Electricity), TimeRAF reports lower zero-shot MSE than the frozen backbone alone and competitive or better results than several TSFM baselines, with ablations over retriever design, integration method, knowledge-base size, and number of candidates.

Significance. If the central empirical claim is valid, this is a practically valuable contribution: a lightweight, frozen-backbone enhancement that turns a general TSFM into a retrieval-augmented forecaster without retraining the foundation model. The method is clearly described, the metric-based retrieval supervision is a legitimate training signal rather than circular reasoning, and the ablation coverage is reasonably broad (retriever variants, Channel Prompting, knowledge-base size, candidate count). The paper does not claim or provide machine-checked proofs, and its value is empirical; at present the empirical evidence is too thin and potentially contaminated to establish the claimed cross-domain generalization.

major comments (3)
  1. [§3.3, §4.1, Appendix A.2] The central 'external knowledge' claim depends on the knowledge base being genuinely external to the evaluation datasets. The paper states that the training subset is drawn from LOTSA and UTSD and that the knowledge base is selected from those same training datasets, but it never states that ETTh1/ETTh2/ETTm1/ETTm2, Weather, or Electricity are excluded from that subset or from the knowledge base. These six datasets are canonical open benchmarks and are very likely members of LOTSA and/or UTSD. Because the inference-time retriever can select sliding-window candidates that are near-duplicates of the historical portion of the test series, the reported gains (e.g., ETTh1 0.359 vs. 0.364; Weather 0.152 vs. 0.158) could reflect test-domain leakage rather than cross-domain retrieval. Table 4's TimeRAFD variant, which deliberately uses the training split of the same dataset as the knowledge base, demonstrates that the mechanism can exploit such same-domain data. The authors must disclose the exact dataset identifiers in the 320M-point training subset and in the knowledge base, confirm whether any evaluation series are present, and rerun the main tables with a knowledge base from which all evaluation-domain series are removed; they should also report how often retrieved candidates come from the same source series as the query.
  2. [Appendix B.2 and Table 2] The experimental protocol does not support the word 'significant' as used in the abstract and Section 1. Appendix B.2 says 'for different test datasets, we reported the best results,' which indicates test-set-based selection without a separate validation protocol, and no standard deviations or number of seeds are reported anywhere. The absolute gains over the frozen backbone are small: 0.359 vs. 0.364 on ETTh1, 0.177 vs. 0.186 on ETTm2, 0.168 vs. 0.170 on Electricity; Table 10 shows that most knowledge-base-size differences appear in the fourth decimal place. The authors should specify the model-selection procedure, report mean and standard deviation over at least three independent training runs, and provide paired significance tests (e.g., Wilcoxon or permutation tests over evaluation windows) for the TimeRAF-versus-backbone comparison.
  3. [§3.5.2, §4.1, Table 2] Please define precisely what 'zero-shot' means for TimeRAF. The TSFM backbone is frozen, but the retriever and Channel Prompting are trained on a 320M-point subset of LOTSA/UTSD with ground-truth labels, and the inference knowledge base is the same as the training knowledge base. This is a legitimate design, but it is not zero-shot in the same sense as the TSFM baselines in Table 2, which perform no gradient updates on any of these data. The current wording in the abstract and Section 1 invites the reader to think the model has not seen the evaluation domains. The authors should either rename the setting (e.g., 'frozen backbone with retrieval adaptation') or explicitly state which components are trained on which corpora, and should add a true held-out-domain experiment (a dataset not present in LOTSA or UTSD) to demonstrate cross-domain retrieval generalization.
minor comments (6)
  1. [§3.2] The phrase 'channel inpendent strategy' should read 'channel independent strategy'.
  2. [§3.5.1] In Equation (5), the notation P_aug is used but never defined; please clarify whether it is computed via Equation (3) using the augmented candidates or using the original candidates.
  3. [§3.5.2] The sentence 'Combined with the loss utilized for updating the retriever,, the whole training loss is...' contains a typo and should be rephrased.
  4. [§3.4] The notation 'ez = {fz1, ..., fzk}' and the use of Avg(MLP(z1), ..., MLP(zk)) should be defined more cleanly; there is also a typo 'to to incorporate' in the same section.
  5. [Appendix C.3] The text says 'As shown in Figure 1' but should refer to Figure 4, and there is a typo 'valuate' in the same passage; the numbers in Table 10 appear to be concatenated without separators (e.g., '0.35920.3598') and need reformatting.
  6. [Figure 3] The y-axis label of Figure 3 is missing, and the '5% Few shot' baseline is described only in the caption; please define it in the main text.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the retriever is supervised by ground-truth forecast quality, the backbone is frozen, and the knowledge base is built from a listed training subset distinct from the evaluation benchmarks.

full rationale

The paper's core claim is empirical: a frozen TTM-Base backbone plus a learned retriever and Channel Prompting improves zero-shot MSE on six benchmarks. The retriever is trained with a supervised signal derived from the forecast metric M(F(x, ci), y) in Eq. 3, which uses ground-truth labels; this is a legitimate training objective, not a circular definition of the output. The forecaster F remains frozen, and the retrieved knowledge is integrated through a separately trained MLP-based Channel Prompting module, so the prediction is not defined in terms of the training target by construction. The knowledge base is constructed from the training subset listed in Table 5, which contains datasets such as BDG-2, Solar Power, Uber TLC, Subseasonal Precipitation, Kaggle Web Traffic, Wiki-Rolling, M5, and MotorImagery, while the evaluation datasets (ETTh1, ETTh2, ETTm1, ETTm2, Weather, Electricity) are described separately in Appendix A.2 and do not appear in the listed training subset. No load-bearing step reduces to a self-citation or to an imported uniqueness theorem, and the acknowledged limitations in the Conclusion are honest heuristic caveats rather than circular reasoning. A possible concern about benchmark membership in LOTSA/UTSD would be a data-leakage or experimental-validity issue, not a circularity of the derivation, and the paper's own dataset table provides no evidence that the evaluation series are included in the knowledge base. Therefore no specific circular reduction can be exhibited from the paper's equations or citations, and the appropriate finding is no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The core assumptions are that external sequences are useful, that metric-trained retrieval transfers to label-free inference, and that standard softmax/KL optimization works with the frozen backbone. The most consequential, unstated assumption is that the evaluation datasets are absent from the LOTSA/UTSD-derived knowledge base; the paper provides no overlap analysis.

free parameters (6)
  • k (number of retrieved candidates) = 8 (default)
    Used during inference; the ablation in Section 4.4.2 shows k=8 works well but there is no principled selection.
  • tau_m (metric temperature) = Not reported in the paper.
    Controls the softmax sharpness in Eq. 3 and directly affects the retriever training signal.
  • tau_s (similarity temperature) = Not reported in the paper.
    Controls the softmax of retrieval scores in Eq. 4 and shapes the KL divergence target.
  • rho (candidate augmentation probability) = Not reported in the paper.
    Introduced in Eq. 5 to replace candidates with random samples; without this value the augmentation procedure is underspecified.
  • lambda (retriever loss weight) = 1
    Weights the augmentation loss in Eq. 7; fixed at 1 with no sensitivity analysis.
  • Knowledge base size and per-domain balance = ~3M points, ~16% per domain
    Hand-curated from training data; Figure 4 shows performance degrades as the KB shrinks, so this choice is load-bearing.
assumptions (4)
  • domain assumption Retrieved sequences from other domains or datasets can provide information useful for forecasting the target series.
    Core premise of retrieval-augmented forecasting; supported only indirectly by the ablations in Table 3.
  • domain assumption A retriever trained with ground-truth-based metric probabilities will generalize at inference to select useful candidates without labels.
    The training signal in Eq. 3 uses the ground-truth y; at inference y is unavailable, and the paper provides no direct evidence of transfer.
  • domain assumption The evaluation datasets are absent from the LOTSA/UTSD-derived training subset and knowledge base.
    Never stated in the paper; if false, the zero-shot evaluation is contaminated because the retriever can access training splits of the test datasets. This is the paper's most consequential unstated premise.
  • standard math Standard softmax, KL divergence, and dot-product operations are valid in this setting and interact correctly with the frozen TTM backbone.
    No analysis of distribution shift in the backbone's embedding space is provided, but the math itself is standard.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TimeRAF: Retrieval-Augmented Foundation model for Zero-shot Time Series Forecasting." pith.science (2026). https://pith.science/paper/QQNBEOD6

@misc{pith2026241220810,
  author       = {Pith},
  title        = {Pith review of: TimeRAF: Retrieval-Augmented Foundation model for Zero-shot Time Series Forecasting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QQNBEOD6}},
  note         = {Machine review of arXiv:2412.20810}
}
read the original abstract

Time series forecasting plays a crucial role in data mining, driving rapid advancements across numerous industries. With the emergence of large models, time series foundation models (TSFMs) have exhibited remarkable generalization capabilities, such as zero-shot learning, through large-scale pre-training. Meanwhile, Retrieval-Augmented Generation (RAG) methods have been widely employed to enhance the performance of foundation models on unseen data, allowing models to access to external knowledge. In this paper, we introduce TimeRAF, a Retrieval-Augmented Forecasting model that enhance zero-shot time series forecasting through retrieval-augmented techniques. We develop customized time series knowledge bases that are tailored to the specific forecasting tasks. TimeRAF employs an end-to-end learnable retriever to extract valuable information from the knowledge base. Additionally, we propose Channel Prompting for knowledge integration, which effectively extracts relevant information from the retrieved knowledge along the channel dimension. Extensive experiments demonstrate the effectiveness of our model, showing significant improvement across various domains and datasets.

Figures

Figures reproduced from arXiv: 2412.20810 by the authors.

Figure 1
Figure 1. Left: Time series foundation models (TSFMs), while capable of zero-shot forecasting, are limited by insufficient prior knowledge, resulting in constrained prediction accuracy. Right: By dynamically retrieving relevant information from an external knowledge base, our TimeRAF enhances prediction accuracy, leading to more precise zero-shot forecasting performance. similar weather conditions—without requiring extensive … view at source ↗
Figure 2
Figure 2. Overview of TimeRAF: TimeRAF utilizes a retriever to dynamically retrieve relevant candidates from an external knowledge base and then utilizes the proposed Channel Prompting module to integrate knowledge between the retrieved data and the input. The knowledge-enhanced embeddings are subsequently fed into the backbone of the foundation model to improve forecasting results. During training, the backbone remains froze… view at source ↗
Figure 3
Figure 3. Improvement by TimeRAF on zero-shot forecasting. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Influence of knowledge base size. Smaller knowledge base provides less information, leading to worse performance. Size of Knowledge Base: The size of knowledge base also plays a vital role in the framework, determining the extent of external knowledge that can be acces…
Figure 5
Figure 5. Figure 5: Influence of the Candidates Number k. As k increases, the performance gradually improves due to the integration of more relevant knowledge. However, when k exceeds a certain threshold, the abundance of information can introduce redundancy, negatively affecting the pred…
Figure 6
Figure 6. Figure 6: Case Study on Retrieved Knowledge. (a) Example A: The retrieved knowledge shares similar periodicity and subtle fluctuations with the input, facilitating the forecaster’s ability to effec￾tively capture the prior knowledge inherent in the input, thereby improving predi…
Figure 7
Figure 7. Figure 7: Visualization of zero-shot forecasting across different datasets. [PITH_FULL_IMAGE:figures/full_fig_p020_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Dynamic Perturbed Adaptive Method for Infinite Task-Conflicting Time Series

    cs.LG 2025-05 reject novelty 4.0 of 10

    A trunk-branch method for adapting to conflicting time series tasks reports large error reductions on a synthetic benchmark, but the comparison is confounded by unequal adaptation budgets and the theory overclaims rel...

  2. Large Language models for Time Series Analysis: Techniques, Applications, and Challenges

    cs.LG 2025-05 reject novelty 3.0 of 10

    A review of LLM-based time series analysis that proposes several taxonomies, but is undermined by citation errors and a lack of systematic methodology.

Reference graph

Works this paper leans on

42 extracted references · 21 canonical work pages · cited by 2 Pith papers

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Gluonts: Probabilistic and neural time series modeling in python

    Alexander Alexandrov, Konstantinos Benidis, Michael Bohlke-Schneider, Valentin Flunkert, Jan Gasthaus, Tim Januschowski, Danielle C Maddix, Syama Rangapuram, David Salinas, Jasper Schulz, et al. Gluonts: Probabilistic and neural time series modeling in python. Journal of Machine Learning Research, 21 0 (116): 0 1--6, 2020

    work page 2020

  3. [3]

    Chronos: Learning the language of time series

    Abdul Fatir Ansari, Lorenzo Stella, Caner Turkmen, Xiyuan Zhang, Pedro Mercado, Huibin Shen, Oleksandr Shchur, Syama Sundar Rangapuram, Sebastian Pineda Arango, Shubham Kapoor, et al. Chronos: Learning the language of time series. arXiv preprint arXiv:2403.07815, 2024

    arXiv 2024

  4. [4]

    Re-imagen: Retrieval-augmented text-to-image generator

    Wenhu Chen, Hexiang Hu, Chitwan Saharia, and William W Cohen. Re-imagen: Retrieval-augmented text-to-image generator. In The Eleventh International Conference on Learning Representations, 2023

    work page 2023

  5. [5]

    A decoder-only foundation model for time-series forecasting

    Abhimanyu Das, Weihao Kong, Rajat Sen, and Yichen Zhou. A decoder-only foundation model for time-series forecasting. In Forty-first International Conference on Machine Learning, 2024

    2024

  6. [6]

    The ucr time series archive

    Hoang Anh Dau, Anthony Bagnall, Kaveh Kamgar, Chin-Chia Michael Yeh, Yan Zhu, Shaghayegh Gharghabi, Chotirat Ann Ratanamahatana, and Eamonn Keogh. The ucr time series archive. IEEE/CAA Journal of Automatica Sinica, pp.\ 1293--1305, 2019

    work page 2019

  7. [7]

    Forecastpfn: Synthetically-trained zero-shot forecasting

    Samuel Dooley, Gurnoor Singh Khurana, Chirag Mohapatra, Siddartha V Naidu, and Colin White. Forecastpfn: Synthetically-trained zero-shot forecasting. Advances in Neural Information Processing Systems, 36, 2024

    work page 2024

  8. [8]

    Tiny time mixers (ttms): Fast pre-trained models for enhanced zero/few-shot forecasting of multivariate time series

    Vijay Ekambaram, Arindam Jati, Pankaj Dayama, Sumanta Mukherjee, Nam H Nguyen, Wesley M Gifford, Chandra Reddy, and Jayant Kalagnanam. Tiny time mixers (ttms): Fast pre-trained models for enhanced zero/few-shot forecasting of multivariate time series. CoRR, 2024

    work page 2024

Show all 42 references
  1. [9]

    Buildingsbench: A large-scale dataset of 900k buildings and benchmark for short-term load forecasting

    Patrick Emami, Abhijeet Sahu, and Peter Graf. Buildingsbench: A large-scale dataset of 900k buildings and benchmark for short-term load forecasting. Advances in Neural Information Processing Systems, 2023

    2023

  2. [10]

    Timegpt-1

    Azul Garza and Max Mergenthaler-Canseco. Timegpt-1. arXiv preprint arXiv:2310.03589, 2023

    2023 arXiv

  3. [11]

    Monash time series forecasting archive

    Rakshitha Godahewa, Christoph Bergmeir, Geoffrey I Webb, Rob J Hyndman, and Pablo Montero-Manso. Monash time series forecasting archive. arXiv preprint arXiv:2105.06643, 2021

    2021 arXiv

  4. [12]

    Moment: A family of open time-series foundation models

    Mononito Goswami, Konrad Szafer, Arjun Choudhry, Yifu Cai, Shuo Li, and Artur Dubrawski. Moment: A family of open time-series foundation models. arXiv preprint arXiv:2402.03885, 2024

    2024 arXiv

  5. [13]

    Learning retrieval augmentation for personalized dialogue generation

    Qiushi Huang, Shuai Fu, Xubo Liu, Wenwu Wang, Tom Ko, Yu Zhang, and Lilian Tang. Learning retrieval augmentation for personalized dialogue generation. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. Association for Computational Lingu...

    2023

  6. [14]

    Towards efficient and comprehensive urban spatial-temporal prediction: A unified library and performance benchmark

    Jiawei Jiang, Chengkai Han, Wenjun Jiang, Wayne Xin Zhao, and Jingyuan Wang. Towards efficient and comprehensive urban spatial-temporal prediction: A unified library and performance benchmark. arXiv e-prints, 2023

    2023

  7. [15]

    Trafficbert: Pre-trained model with large-scale data for long-range traffic flow forecasting

    KyoHoon Jin, JeongA Wi, EunJu Lee, ShinJin Kang, SooKyun Kim, and YoungBin Kim. Trafficbert: Pre-trained model with large-scale data for long-range traffic flow forecasting. Expert Systems with Applications, 186: 0 115738, 2021

    2021

  8. [16]

    Time-llm: Time series forecasting by reprogramming large language models

    Ming Jin, Shiyu Wang, Lintao Ma, Zhixuan Chu, James Y Zhang, Xiaoming Shi, Pin-Yu Chen, Yuxuan Liang, Yuan-Fang Li, Shirui Pan, et al. Time-llm: Time series forecasting by reprogramming large language models. In The Twelfth International Conference on Learning Representations, 2023

    2023

  9. [17]

    Retrieval based time series forecasting

    Baoyu Jing, Si Zhang, Yada Zhu, Bin Peng, Kaiyu Guan, Andrew Margenot, and Hanghang Tong. Retrieval based time series forecasting. arXiv preprint arXiv:2209.13525, 2022

    2022 arXiv

  10. [18]

    Dense passage retrieval for open-domain question answering

    Vladimir Karpukhin, Barlas O g uz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. Dense passage retrieval for open-domain question answering. arXiv preprint arXiv:2004.04906, 2020

    2004 arXiv

  11. [19]

    Frozen language model helps ecg zero-shot learning

    Jun Li, Che Liu, Sibo Cheng, Rossella Arcucci, and Shenda Hong. Frozen language model helps ecg zero-shot learning. In Medical Imaging with Deep Learning, pp.\ 402--415. PMLR, 2024

    2024

  12. [20]

    Foundation models for time series analysis: A tutorial and survey

    Yuxuan Liang, Haomin Wen, Yuqi Nie, Yushan Jiang, Ming Jin, Dongjin Song, Shirui Pan, and Qingsong Wen. Foundation models for time series analysis: A tutorial and survey. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp.\ 6555--6565, 2024

    2024

  13. [21]

    Learning customized visual models with retrieval-augmented knowledge

    Haotian Liu, Kilho Son, Jianwei Yang, Ce Liu, Jianfeng Gao, Yong Jae Lee, and Chunyuan Li. Learning customized visual models with retrieval-augmented knowledge. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 15148--15158, 2023

    2023

  14. [22]

    itransformer: Inverted transformers are effective for time series forecasting

    Yong Liu, Tengge Hu, Haoran Zhang, Haixu Wu, Shiyu Wang, Lintao Ma, and Mingsheng Long. itransformer: Inverted transformers are effective for time series forecasting. In The Twelfth International Conference on Learning Representations

  15. [23]

    Timer: Generative pre-trained transformers are large time series models

    Yong Liu, Haoran Zhang, Chenyu Li, Xiangdong Huang, Jianmin Wang, and Mingsheng Long. Timer: Generative pre-trained transformers are large time series models. In Forty-first International Conference on Machine Learning, 2024

    2024

  16. [24]

    Subseasonalclimateusa: a dataset for subseasonal forecasting and benchmarking

    Soukayna Mouatadid, Paulo Orenstein, Genevieve Flaspohler, Miruna Oprescu, Judah Cohen, Franklyn Wang, Sean Knight, Maria Geogdzhayeva, Sam Levang, Ernest Fraenkel, et al. Subseasonalclimateusa: a dataset for subseasonal forecasting and benchmarking. Advances in Neural Informa...

    2024

  17. [25]

    A time series is worth 64 words: Long-term forecasting with transformers

    Yuqi Nie, Nam H Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. A time series is worth 64 words: Long-term forecasting with transformers. In The Eleventh International Conference on Learning Representations, 2023

    2023

  18. [26]

    Lag-llama: Towards foundation models for time series forecasting

    Kashif Rasul, Arjun Ashok, Andrew Robert Williams, Arian Khorasani, George Adamopoulos, Rishika Bhagwatkar, Marin Bilo s , Hena Ghonia, Nadhir Vincent Hassen, Anderson Schneider, et al. Lag-llama: Towards foundation models for time series forecasting. arXiv preprint arXiv:2310...

    arXiv

  19. [27]

    arrow: Integration to 'Apache' 'Arrow', 2024

    Neal Richardson, Ian Cook, Nic Crane, Dewey Dunnington, Romain François, Jonathan Keane, Dragoș Moldovan-Grünfeld, Jeroen Ooms, Jacob Wujciak-Jens, and Apache Arrow . arrow: Integration to 'Apache' 'Arrow', 2024. URL https://github.com/apache/arrow/. R package version 17.0.0, ...

    2024

  20. [28]

    Replug: Retrieval-augmented black-box language models

    Weijia Shi, Sewon Min, Michihiro Yasunaga, Minjoon Seo, Rich James, Mike Lewis, Luke Zettlemoyer, and Wen-tau Yih. Replug: Retrieval-augmented black-box language models. arXiv preprint arXiv:2301.12652, 2023

    2023 arXiv

  21. [29]

    Llama: Open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timoth \'e e Lacroix, Baptiste Rozi \`e re, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023

    2023 arXiv

  22. [30]

    Ratsf: Empowering customer service volume management through retrieval-augmented time-series forecasting

    Tianfeng Wang and Gaojie Cui. Ratsf: Empowering customer service volume management through retrieval-augmented time-series forecasting. arXiv preprint arXiv:2403.04180, 2024

    2024 arXiv

  23. [31]

    Unified training of universal time series forecasting transformers

    Gerald Woo, Chenghao Liu, Akshat Kumar, Caiming Xiong, Silvio Savarese, and Doyen Sahoo. Unified training of universal time series forecasting transformers. arXiv preprint arXiv:2402.02592, 2024

    2024 arXiv

  24. [32]

    Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting

    Haixu Wu, Jiehui Xu, Jianmin Wang, and Mingsheng Long. Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting. Advances in neural information processing systems, 34: 0 22419--22430, 2021

    2021

  25. [33]

    Timesnet: Temporal 2d-variation modeling for general time series analysis

    Haixu Wu, Tengge Hu, Yong Liu, Hang Zhou, Jianmin Wang, and Mingsheng Long. Timesnet: Temporal 2d-variation modeling for general time series analysis. In The Eleventh International Conference on Learning Representations, 2023 a

    2023

  26. [34]

    Interpretable weather forecasting for worldwide stations with a unified deep model

    Haixu Wu, Hang Zhou, Mingsheng Long, and Jianmin Wang. Interpretable weather forecasting for worldwide stations with a unified deep model. Nature Machine Intelligence, 5 0 (6): 0 602--611, 2023 b

    2023

  27. [35]

    Temporal data meets llm--explainable financial time series forecasting

    Xinli Yu, Zheng Chen, Yuan Ling, Shujing Dong, Zongyi Liu, and Yanbin Lu. Temporal data meets llm--explainable financial time series forecasting. arXiv preprint arXiv:2306.11025, 2023

    2023 arXiv

  28. [36]

    Are transformers effective for time series forecasting? In Proceedings of the AAAI conference on artificial intelligence, 2023

    Ailing Zeng, Muxi Chen, Lei Zhang, and Qiang Xu. Are transformers effective for time series forecasting? In Proceedings of the AAAI conference on artificial intelligence, 2023

    2023

  29. [37]

    Informer: Beyond efficient transformer for long sequence time-series forecasting

    Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang. Informer: Beyond efficient transformer for long sequence time-series forecasting. In Proceedings of the AAAI conference on artificial intelligence, 2021

    2021

  30. [38]

    Doccoder: Generating code by retrieving and reading docs

    Shuyan Zhou, Uri Alon, Frank F Xu, Zhengbao JIang, and Graham Neubig. Doccoder: Generating code by retrieving and reading docs. arXiv preprint arXiv:2207.05987, 2022

    2022 arXiv

  31. [39]

    One fits all: Power general time series analysis by pretrained lm

    Tian Zhou, Peisong Niu, Liang Sun, Rong Jin, et al. One fits all: Power general time series analysis by pretrained lm. Advances in neural information processing systems, 2023

    2023

  32. [40]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  33. [41]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  34. [42]

    ׀H 0i _

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    1980

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.