Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

MoTime: A Dataset Suite for Multimodal Time Series Forecasting

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read MoTime benchmark: text and image context improve forecasts across 869,466 series, including cold-start cases.

desk verdict A genuinely useful multimodal forecasting suite, but the cold-start claim rests on scraped text that hasn't been checked for future leakage, and the empirical support for modality gains is thinner than the abstract suggests. read the letter →

arxiv 2505.15072 v2 pith:PX6I7WKD submitted 2025-05-21 cs.LG cs.CLcs.DBcs.IR

Xin Zhou , Weiqing Wang , Francisco J. Baldán , Wray Buntine , Christoph Bergmeir This is my paper
classification cs.LGcs.CLcs.DBcs.IR
keywords multimodaltimeseriesforecastingdatasetsuitecold-startvarying-historyentity-centricmodalitiestext-conditionedimage-conditionedretrieval-augmentedgeneration
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

MoTime is a dataset suite that pairs 869,466 item-level time series across eight domains with static external modalities—text descriptions, metadata, and images—and defines two evaluation protocols: varying-history forecasting and cold-start forecasting where no history exists. The paper's central claim is that external modalities can improve forecasting performance in both scenarios, with the clearest gains on short or sparse series where the temporal signal alone is weak. This matters because most forecasting benchmarks are unimodal, so there has been no large-scale, reusable way to test whether contextual information such as an item's category, description, or image helps a model anticipate popularity dynamics. The suite is built by re-purposing recommender-system and web-traffic sources into item-centric series, aligned to modalities through consistent identifiers, and is released publicly so the field can evaluate context-aware forecasting systematically.

What carries the argument

The load-bearing object is the MoTime suite itself: eight datasets transformed into item-centric daily or high-frequency popularity series with aligned static modalities, built by aggregating user-item interactions (PixelRec, TaobaoFashion, AmazonReview, Tianchi, MovieLens) or repurposing web and news traffic logs (News, WikiPeople). The two evaluation scenarios are the second piece of machinery: a varying-history protocol that splits training series into long and short groups to measure marginal modality utility, and a cold-start protocol that masks all but the first valid observation so forecasts must come from external signals alone. For cold-start, the mechanism is a retrieval-augmented generation pipeline: frozen LLM embeddings, cosine similarity over a retrieval base, top-k neighbor series and metadata composed into a structured prompt, and an LLM generating the forecast. The two proposed model variants, MultiPatchTST and MultiWPMixer, use a time series encoder plus a frozen LLM encoder to fuse modalities, with images converted to captions before encoding.

What would settle it

Take the MovieLens and WikiPeople splits, crawl the same text sources again, and check whether any summary references events dated after the forecast window (for example, a Wikipedia summary mentioning a person's death year after the series ends). Re-run the cold-start and short-history experiments with those sentences redacted; if the multimodal advantage shrinks or disappears, the reported gains are at least partly leakage rather than genuine use of context.

Watch

Extended reading notes

Core claim

The paper introduces MoTime as the largest publicly available multimodal time series forecasting suite of its kind, covering 869,466 series with text, image, or metadata alignment and reusable protocols. In varying-history forecasting, it adapts PatchTST and WPMixer into multimodal variants by concatenating a time series encoder's patch embeddings with frozen-LLM text (and image-caption) embeddings; in cold-start forecasting, it uses a retrieval-augmented pipeline where a target entity's text description retrieves the top-k most similar entities and an LLM generates the forecast from their series and metadata. The empirical finding is that external modalities generally improve performance in both scenarios, but the benefit is dataset- and horizon-dependent: gains are notable for short series on some datasets, and cold-start retrieval beats a simple average baseline on most of the six tested datasets. The authors frame the contribution as infrastructure: a reusable suite and scenario-driven evaluation rather than a new forecasting model.

Load-bearing premise

The externally scraped text for MovieLens and WikiPeople—movie overviews and Wikipedia summaries—is assumed to contain no information from after the forecast period, but the paper performs no timestamp or content-revision audit, so a summary mentioning later events (such as a person's death or a film's later reception) could hand the model future knowledge and inflate the reported cold-start and short-history gains.

Editorial extensions

If this is right

  • If the central claim holds, cold-start forecasting—predicting an item's popularity before any observations exist—becomes a tractable, benchmarkable task rather than a niche setting.
  • Static entity-level text and images can partially substitute for missing history, which would let recommender-system logs be reused as forecasting benchmarks wherever metadata already exists.
  • Because modality gains vary with history length, horizon, and sparsity, future benchmarks should report short/long and horizon-specific scores rather than a single aggregate, or they will hide where context actually helps.
  • The 869,466-series scale gives multimodal forecasting a training and evaluation corpus large enough to support large-model experiments analogous to those in unimodal forecasting.

Reading between the lines

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

  • I read the paper as evidence that a series' identity—what the object is—is a learnable prior that matters most exactly when the observed signal is thinnest; a testable extension would be ablating modality type per dataset to see which channel carries the cold-start signal.
  • The retrieval-augmented cold-start design is effectively nearest-neighbour forecasting in an embedding space; a direct implication is that retrieval quality, not generation, sets the ceiling, so a similarity-weighted average over retrieved series might close much of the gap to LLM generation on datasets where the two already match.
  • The authors acknowledge that short series were not upsampled and sparsity was not systematically varied; a natural follow-up experiment would control sparsity and history length independently to separate their contributions to modality utility.
  • Because text summaries for MovieLens and WikiPeople were crawled from current public sources without a timestamp audit, the reported cold-start gains could be inflated by post-forecast information; the suite should ship with crawl dates or content-revision dates so users can audit leakage.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper introduces MoTime, a suite of eight multimodal time-series forecasting datasets constructed by re-purposing recommender-system and web-popularity data. Each dataset pairs time series with static text, image, and/or metadata modalities, and the suite is released publicly. The authors define two evaluation scenarios: varying-history forecasting, where short and long historical windows are contrasted, and cold-start forecasting, where no target history is available and predictions are made from textual descriptions, retrieved similar series, and GPT-4o-mini generation. Experiments with adapted PatchTST and WPMixer variants, plus LLM-based cold-start forecasts, are used to argue that external modalities can improve forecasting in both scenarios, with gains that vary by dataset, sparsity, and horizon.

Significance. If the empirical claims hold, MoTime would be a valuable infrastructure contribution: it is substantially larger than existing multimodal time-series benchmarks (869,466 series across eight domains), includes static entity-level text/image/metadata alignments rather than only dynamic event streams, and explicitly operationalizes cold-start forecasting, which is rarely benchmarked. The paper is also transparent about many construction details and limitations, and the data release is a concrete asset for the community. However, the central claim that external modalities improve cold-start forecasting is currently supported by experiments that lack a leakage audit and statistical grounding, so the significance is conditional on those issues being resolved.

major comments (4)
  1. [§3.2, §4.2, Fig. 2] The cold-start modality-utility claim is exposed to temporal leakage from externally scraped text. Section 3.2 states that for MovieLens and WikiPeople text is obtained by crawling movie metadata or Wikipedia summaries and linked by IDs, but no timestamp check is described. WikiPeople series end in mid-2017 (the Kaggle competition), while Wikipedia summaries scraped circa 2025 can contain post-2017 facts such as deaths, elections, or later career events. The cold-start protocol in Section 4.2 removes all but the first valid observation but feeds the target's current text description into retrieval and GPT generation; the retrieved-series averaging baseline does not receive this semantic knowledge. This violates the no-history premise and can inflate the GPT-versus-average gap in Figure 2. AmazonReview has a parallel issue: aggregate rating metadata computed over the full 2013--2023 period post-dates the cold-start forecast points. The authors should either use timestamped text snapshots (e.g., Wikipedia revision dates, aggregated metadata computed only up to each prediction point) or demonstrate empirically that the reported cold-start results are unchanged under a strict leakage audit.
  2. [§5.1, Tables 3 and 15] All headline comparisons are reported without confidence intervals, standard deviations across seeds, or significance tests. Many differences are at the third decimal (e.g., Table 3, Amazon horizon 1: MultiPatchTST 0.373 vs DLinear 0.373, with bold/underline assigned by full precision). The large number of series per dataset could make standard errors very small, but that is precisely why significance tests, effect-size reporting, or seed variance are needed to distinguish real gains from noise. Moreover, comparing PatchTST with MultiPatchTST and WPMixer with MultiWPMixer changes both the architecture and the input modality; a genuine modality-utility estimate requires a controlled ablation with the same architecture and the text/image input dropped, or a statistical interaction test.
  3. [§4.1, §6, Table 15] The short-series benefit claim is confounded by training imbalance. Section 4.1 trains jointly on a 1:1 split of short and long series with no upsampling, and Section 6 acknowledges that short series may contribute less to gradient updates. Since the abstract claims 'particularly strong benefits for short series in some datasets,' the observation that MovieLens short series improve with MultiWPMixer (Section 5.2.1) could be an artifact of underfitting short series rather than a modality effect. The authors should rebalance the training data (e.g., oversample short windows or add a weighting term) and rerun, or at least report a sensitivity analysis varying the short/long ratio and the horizon thresholds listed in Section 5.1.
  4. [§5.2.2, §5.1] Cold-start evaluation uses a fixed random sample of 30 series per dataset and a single top-4 retrieval setting. With sparse, bursty series (e.g., PixelRec and Tianchi), single-split results on such a small sample are unstable, especially when WRMSPE denominators are near zero. The paper reports no variance across sampling seeds and no sensitivity to the retrieval top-k, the cold-start sample size, or the PixelRec density-0.4 and MovieLens 1,000-series filters introduced in Section 5.1. I therefore do not trust the specific magnitudes in Figure 2, even if the qualitative direction may be right.
minor comments (6)
  1. [Abstract, §5.2.1] The abstract says external modalities give 'particularly strong benefits for short series in some datasets,' but the main text concludes the short-series hypothesis holds only in specific cases and that performance trends are largely consistent across short and long subsets; the abstract should be softened to match this nuance.
  2. [§3.1, Table 2, §3.2] VISUELLE is listed in Table 2 and Section 3.1 but is not processed or experimented on; the main text should state clearly whether VISUELLE is included in the released suite and, if so, why it appears in the statistics table. Also, Section 3.1 says 'The time series is item sales' for VISUELLE, which appears inconsistent with the engagement-statistics description.
  3. [§3.1, §5.1, §8.1.6] The News dataset evaluation uses a filtered subset (topic 'obama', platform Facebook), which is described only in Appendix 8.1.6; this filtering decision is important for reproducibility and should be stated in the main text, ideally in Section 3.1 or Section 5.1.
  4. [§3.3, Table 3, Table 13] There are several presentation issues: 'obervations' in Section 3.3, inconsistent spellings 'Movielens' vs 'MovieLens', unformatted WikiPeople RMSE values such as '21977' in Table 3, and the placeholder 'gragality' in the prompt template of Table 13.
  5. [Tables 7--10] The AmazonReview per-category text statistics show identical minimum, maximum, mean, and median values for many categories (e.g., char_len_max=3938 for many categories), which suggests the statistics describe a single long concatenated metadata string; please clarify how these text fields are constructed and whether truncation is applied before embedding.
  6. [Fig. 2] Figure 2 reports cold-start results only as bars without numeric values or error bars; please add an appendix table with the exact RMSE/WRMSPE values and repeated-seed variation so readers can assess the magnitude of the GPT-versus-average differences.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the paper's claims are empirical benchmark comparisons, and its self-citations are contextual rather than load-bearing.

full rationale

MoTime is a dataset-and-benchmark paper whose central claims are empirical: that the suite enables multimodal forecasting evaluation and that external modalities help in varying-history and cold-start settings. These claims are supported by held-out test evaluations of concrete models (DLinear, PatchTST, WPMixer, MultiPatchTST, MultiWPMixer, and a GPT-4o-mini retrieval pipeline) against baselines, with no parameter fitted to the reported test outcomes. The self-citations to TextFusionHTS [63], the CCWTF codebase [61], and reindexing [62] are used as architectural or implementation inspiration, not as evidence of correctness or as a uniqueness argument, so they are not load-bearing. The reader-identified temporal-leakage risk in scraped Wikipedia summaries and movie metadata is a data-validity or benchmark-design concern, not circularity: it does not make any predicted quantity equal to an input by construction. No equation, definition, or fitted parameter reduces to itself, and no 'prediction' is a renamed fit. Therefore the appropriate finding is no significant circularity.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claims rest on hand-chosen experimental thresholds and unvalidated assumptions about modality timing and retrieval relevance. These are testable, but not tested in the paper.

free parameters (5)
  • short/long history thresholds = 100/50/20/18 steps per dataset
    Chosen cutoffs in Section 5.1 define which series count as 'short'; the main moderation conclusion (short series benefit more from modalities) depends on these cutoffs.
  • retrieval top-k = 4
    Cold-start GPT pipeline retrieves top 4 similar series; changing k likely changes forecast quality, and no sensitivity analysis is reported.
  • cold-start sample size = 30 series per dataset
    Only 30 random channels are used as cold-start targets, with no multiple seeds or variance reporting.
  • PixelRec density filter = density >= 0.4
    Series with density below 0.4 are removed for experiments, altering the evaluated distribution; the threshold is not justified.
  • MovieLens sample size = 1000 series
    Randomly sampled 1000 from 10,505 series; the sampled subset may not represent the full dataset's sparsity.
assumptions (4)
  • domain assumption Static entity-level modalities are time-invariant and available at prediction time without future information.
    The cold-start and varying-history protocols feed text and images into the model as if they were known before the forecast horizon; this is not validated, and scraped summaries may contain post-forecast facts.
  • domain assumption Aggregated user interaction counts (views, purchases, reviews) constitute valid forecasting targets comparable across entities.
    Recommender logs are re-purposed into popularity series; no analysis of the resulting target's stationarity or comparability is provided.
  • domain assumption Semantic similarity of LLM embeddings of text is a good proxy for similarity of future time series behavior.
    The cold-start retrieval pipeline assumes that entities with similar descriptions have similar future trajectories; this is the core mechanism but is not evaluated against alternative retrieval signals.
  • domain assumption The 7:1:2 temporal split prevents train/test leakage for all series.
    The split is based on the longest series, but for series of varying lengths and aligned modalities, the effective test ratio and overlap are not fully specified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MoTime: A Dataset Suite for Multimodal Time Series Forecasting." pith.science (2026). https://pith.science/paper/PX6I7WKD

@misc{pith2026250515072,
  author       = {Pith},
  title        = {Pith review of: MoTime: A Dataset Suite for Multimodal Time Series Forecasting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PX6I7WKD}},
  note         = {Machine review of arXiv:2505.15072}
}
read the original abstract

While multimodal data sources are increasingly available from real-world forecasting, most existing research remains on unimodal time series. In this work, we present MoTime, a suite of multimodal time series forecasting datasets that pair temporal signals with external modalities such as text, metadata, and images. Covering diverse domains, MoTime supports structured evaluation of modality utility under two scenarios: 1) the common forecasting task, where varying-length history is available, and 2) cold-start forecasting, where no historical data is available. Experiments show that external modalities can improve forecasting performance in both scenarios, with particularly strong benefits for short series in some datasets, though the impact varies depending on data characteristics. By making datasets and findings publicly available, we aim to support more comprehensive and realistic benchmarks in future multimodal time series forecasting research.

Figures

Figures reproduced from arXiv: 2505.15072 by the authors.

Figure 1
Figure 1. Monthly demand of a seasonal product, portable fan (orange curve), and a trend-sensitive [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Cold-start forecasting results. Top: RMSE; Bottom: WRMSPE. The yellow bar is the [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. A Survey of Reasoning and Agentic Systems in Time Series with Large Language Models

    cs.AI 2025-09 conditional novelty 5.0 of 10

    The authors organize LLM-based time series reasoning into three exclusive topologies (direct, chain, branch) crossed with four objectives, and use them to label 125 papers, benchmarks, and resources.

Reference graph

Works this paper leans on

64 extracted references · 47 canonical work pages · cited by 1 Pith paper

  1. [1]

    https://en.wikipedia.org/wiki/Dow_Jones_ Industrial_Average

    Dow jones industrial average (djia). https://en.wikipedia.org/wiki/Dow_Jones_ Industrial_Average. Accessed: 2025-05-13

    work page 2025

  2. [2]

    https://www.kaggle.com/competitions/ web-traffic-time-series-forecasting

    Web traffic time series forecasting. https://www.kaggle.com/competitions/ web-traffic-time-series-forecasting. Accessed: 2025-05-11

    work page 2025

  3. [3]

    Dreamdiffusion: Generating high-quality images from brain eeg signals

    Yunpeng Bai et al. Dreamdiffusion: Generating high-quality images from brain eeg signals. arXiv preprint arXiv:2306.16934, 2023

    arXiv 2023

  4. [4]

    Multi-modal financial time-series retrieval through latent space projections

    Tom Bamford, Andrea Coletta, Elizabeth Fons, Sriram Gopalakrishnan, Svitlana Vyetrenko, Tucker Balch, and Manuela Veloso. Multi-modal financial time-series retrieval through latent space projections. InProceedings of the 4th ACM International Conference on AI in Finance (ICAIF ’23), pages 498–506, Brooklyn, NY , USA, November 2023. ACM

    work page 2023

  5. [5]

    TEMPO: Prompt-based generative pre-trained transformer for time series forecasting

    Defu Cao, Furong Jia, Sercan O Arik, Tomas Pfister, Yixiang Zheng, Wen Ye, and Yan Liu. TEMPO: Prompt-based generative pre-trained transformer for time series forecasting. InThe Twelfth International Conference on Learning Representations, 2024

    2024

  6. [6]

    Fashion trend forecasting using machine learning tech- niques: A review

    Audrey Aurelia Chang, Cynthia, Devita, Jihad Fahri Ramadhan, Zainina Kyla Shafira Adnan, Bayu Kanigoro, and Edy Irwansyah. Fashion trend forecasting using machine learning tech- niques: A review. In Radek Silhavy, Petr Silhavy, and Zdenka Prokopova, editors,Data Science and Intelligent Systems, volume 231 ofLecture Notes in Networks and Systems, pages 34–...

    work page 2021

  7. [7]

    LLM4TS: Aligning pre- trained LLMs as data-efficient time-series forecasters.arXiv preprint arXiv:2308.08469, 2023

    Ching Chang, Wei-Yao Wang, Wen-Chih Peng, and Tien-Fu Chen. LLM4TS: Aligning pre- trained LLMs as data-efficient time-series forecasters.arXiv preprint arXiv:2308.08469, 2023

    arXiv 2023

  8. [8]

    Mtbench: A multimodal time series benchmark for temporal reasoning and question answering, 2025

    Jialin Chen, Aosong Feng, Ziyu Zhao, Juan Garza, Gaukhar Nurbek, Cheng Qin, Ali Maatouk, Leandros Tassiulas, Yifeng Gao, and Rex Ying. Mtbench: A multimodal time series benchmark for temporal reasoning and question answering, 2025

    work page 2025

Show all 64 references
  1. [9]

    Terra: A multimodal spatio-temporal dataset spanning the earth

    Wei Chen, Xixuan Hao, Yuankai Wu, and Yuxuan Liang. Terra: A multimodal spatio-temporal dataset spanning the earth. InAdvances in Neural Information Processing Systems 37 (NeurIPS 2024), Datasets and Benchmarks Track, 2024

    2024

  2. [10]

    An image dataset for benchmarking recommender systems with raw pixels.arXiv preprint arXiv:2309.06789, 2023

    Yu Cheng, Yunzhu Pan, Jiaqi Zhang, Yongxin Ni, Aixin Sun, and Fajie Yuan. An image dataset for benchmarking recommender systems with raw pixels.arXiv preprint arXiv:2309.06789, 2023

    2023 arXiv

  3. [11]

    Toto: Time series optimized transformer for observability

    Ben Cohen, Emaad Khwaja, Kan Wang, Charles Masson, Elise Ramé, Youssef Doubli, and Othmane Abou-Amal. Toto: Time series optimized transformer for observability. Technical report, Datadog, 2024. arXiv:2407.07874

    2024 arXiv

  4. [12]

    Nyc bike sharing network: Time-series enhanced nodes and edges dataset, 2024

    NYC Bike Sharing System Contributors. Nyc bike sharing network: Time-series enhanced nodes and edges dataset, 2024

    2024

  5. [13]

    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. InProceedings of the 41st International Conference on Machine Learning (ICML 2024), volume 235 ofProceedings of Machine Learning Research, pages 10148–10167. PM...

    2024

  6. [14]

    Fnspid: A comprehensive financial news dataset in time series.arXiv preprint arXiv:2402.06698, 2024

    Zihan Dong et al. Fnspid: A comprehensive financial news dataset in time series.arXiv preprint arXiv:2402.06698, 2024

    2024 arXiv

  7. [15]

    Uci machine learning repository

    Dheeru Dua and Casey Graff. Uci machine learning repository. https://archive.ics.uci. edu/ml/, 2017. University of California, Irvine, School of Information and Computer Sciences

    2017

  8. [16]

    Nguyen, Wesley M

    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.arXiv preprint arXiv:240...

    2024 arXiv

  9. [17]

    Timegpt: The first foundation model for time series

    Azul Garza, Cristian Challu, and Max Mergenthaler-Canseco. Timegpt: The first foundation model for time series. InTime Series Forecasting Using Generative AI, pages 169–182. Springer, 2025

    2025

  10. [18]

    Webb, Rob J

    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

  11. [19]

    Bridging language and items for retrieval and recommendation.arXiv preprint arXiv:2403.03952, 2024

    Yupeng Hou, Jiacheng Li, Zhankui He, An Yan, Xiusi Chen, and Julian McAuley. Bridging language and items for retrieval and recommendation.arXiv preprint arXiv:2403.03952, 2024

    2024 arXiv

  12. [20]

    Icbhi 2017 challenge: Respiratory sound database

    ICBHI. Icbhi 2017 challenge: Respiratory sound database. https://bhichallenge.med. auth.gr/ICBHI_2017_Challenge, 2017. Accessed: 2025-05-15

    2017

  13. [21]

    GPT4MTS: Prompt-based large language model for multimodal time-series forecasting

    Furong Jia, Kexin Wang, Yixiang Zheng, Defu Cao, and Yan Liu. GPT4MTS: Prompt-based large language model for multimodal time-series forecasting. InProceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 23343–23351, 2024

    2024

  14. [22]

    Zhang, Xiaoming Shi, Pin-Yu Chen, Yuxuan Liang, Yuan-Fang Li, Shirui Pan, and Qingsong Wen

    Ming Jin, Shiyu Wang, Lintao Ma, Zhixuan Chu, James Y . Zhang, Xiaoming Shi, Pin-Yu Chen, Yuxuan Liang, Yuan-Fang Li, Shirui Pan, and Qingsong Wen. Time-llm: Time series forecasting by reprogramming large language models.arXiv preprint arXiv:2310.01728, 2023

    arXiv

  15. [23]

    Mimic-iv, a freely accessible electronic health record dataset.Scientific data, 10(1):1–8, 2023

    Alistair EW Johnson, Tom J Pollard, Nathaniel R Greenbaum, Matthew P Lungren, Chih-ying Deng, Yikuan Peng, Zhi Lu, and Roger G Mark. Mimic-iv, a freely accessible electronic health record dataset.Scientific data, 10(1):1–8, 2023

    2023

  16. [24]

    Mimic-iii, a freely accessible critical care database.Scientific data, 3:160035, 2016

    Alistair EW Johnson, Tom J Pollard, Lu Shen, Li-wei H Lehman, Mengling Feng, Mohammad Ghassemi, Benjamin Moody, Peter Szolovits, Leo Anthony Celi, and Roger G Mark. Mimic-iii, a freely accessible critical care database.Scientific data, 3:160035, 2016

    2016

  17. [25]

    Gdelt: Global data on events, location, and tone.ISA Annual Convention, 2013

    Kalev Leetaru and Philip A Schrodt. Gdelt: Global data on events, location, and tone.ISA Annual Convention, 2013

    2013

  18. [26]

    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. InProceedings of the Medical Imaging with Deep Learning (MIDL) 2023, volume 227 ofProceedings of Machine Learning Research, pages 402–415. PMLR, 2023

    2023

  19. [27]

    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 (KDD ’24), pages 6...

    2024

  20. [28]

    Sasanur, Megha Sharma, Jiaming Cui, Qingsong Wen, Chao Zhang, and B

    Haoxin Liu, Shangqing Xu, Zhiyuan Zhao, Lingkai Kong, Harshavardhan Kamarthi, Aditya B. Sasanur, Megha Sharma, Jiaming Cui, Qingsong Wen, Chao Zhang, and B. Aditya Prakash. Time-mmd: Multi-domain multimodal dataset for time series analysis. InNeurIPS Datasets and Benchmarks Tr...

    2024

  21. [29]

    Aditya Prakash

    Haoxin Liu, Zhiyuan Zhao, Jindong Wang, Harshavardhan Kamarthi, and B. Aditya Prakash. Lstprompt: Large language models as zero-shot time series forecasters by long-short-term prompting. InFindings of the Association for Computational Linguistics: ACL 2024, pages 7832–7840, Ba...

    2024

  22. [30]

    Moirai-moe: Empowering time series foundation models with sparse mixture of experts

    Xu Liu, Juncheng Liu, Gerald Woo, Taha Aksu, Yuxuan Liang, Roger Zimmermann, Chenghao Liu, Silvio Savarese, Caiming Xiong, and Doyen Sahoo. Moirai-moe: Empowering time series foundation models with sparse mixture of experts. InProceedings of the 13th International Conference o...

    2025

  23. [31]

    The m3-competition: results, conclusions and implica- tions.International Journal of Forecasting, 16(4):451–476, 2000

    Spyros Makridakis and Michele Hibon. The m3-competition: results, conclusions and implica- tions.International Journal of Forecasting, 16(4):451–476, 2000

    2000

  24. [32]

    The m4 competition: Results, findings, conclusion and way forward.International Journal of Forecasting, 34(4):802– 808, 2018

    Spyros Makridakis, Evangelos Spiliotis, and Vassilios Assimakopoulos. The m4 competition: Results, findings, conclusion and way forward.International Journal of Forecasting, 34(4):802– 808, 2018. 21

    2018

  25. [33]

    M5 accuracy competi- tion: Results, findings and conclusions.International Journal of Forecasting, 38(4):1279–1282, 2022

    Spyros Makridakis, Evangelos Spiliotis, and Vassilios Assimakopoulos. M5 accuracy competi- tion: Results, findings and conclusions.International Journal of Forecasting, 38(4):1279–1282, 2022

    2022

  26. [34]

    Multi-source social feedback of online news feeds, 2018

    Nuno Moniz and Luís Torgo. Multi-source social feedback of online news feeds, 2018

    2018

  27. [35]

    Wpmixer: Efficient multi-resolution mixing for long-term time series forecasting.arXiv preprint arXiv:2412.17176, 2024

    Md Mahmuddun Nabi Murad et al. Wpmixer: Efficient multi-resolution mixing for long-term time series forecasting.arXiv preprint arXiv:2412.17176, 2024

    2024 arXiv

  28. [36]

    Taxi and Limousine Commission

    New York (N.Y .). Taxi and Limousine Commission. New york city taxi trip data, 2009-2018 (icpsr 37254), 2019

    2009

  29. [37]

    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. InInternational Conference on Learning Representations (ICLR), 2023

    2023

  30. [38]

    Nixtla foundation benchmark.Nixtla, 2024

    Nixtla Inc. Nixtla foundation benchmark.Nixtla, 2024

    2024

  31. [39]

    Papadopoulos, Christos Koutlis, Symeon Papadopoulos, and Ioannis Kompatsiaris

    Stefanos I. Papadopoulos, Christos Koutlis, Symeon Papadopoulos, and Ioannis Kompatsiaris. Multimodal quasi-autoregression: Forecasting the visual popularity of new fashion products. International Journal of Multimedia Information Retrieval, 11(4):717–729, October 2022

    2022

  32. [40]

    Lag-llama: Towards foundation models for probabilistic time series forecasting

    Kashif Rasul, Arjun Ashok, Andrew Robert Williams, Hena Ghonia, Rishika Bhagwatkar, Arian Khorasani, Mohammad Javad Darvishi Bayazi, George Adamopoulos, Roland Riachi, Nadhir Hassen, Marin Biloš, Sahil Garg, Anderson Schneider, Nicolas Chapados, Alexandre Drouin, Valentina Zan...

    2023

  33. [41]

    Multimodal multi-task financial risk forecasting

    Ramit Sawhney, Puneet Mathur, Ayush Mangal, Piyush Khanna, Rajiv Ratn Shah, and Roger Zimmermann. Multimodal multi-task financial risk forecasting. InProceedings of the 28th ACM International Conference on Multimedia (MM ’20), pages 456–465, Seattle, WA, USA, October 2020. ACM

    2020

  34. [42]

    Chronos: Learning the language of time series.arXiv preprint arXiv:2403.07815, 2024

    Amazon Science. Chronos: Learning the language of time series.arXiv preprint arXiv:2403.07815, 2024

    2024 arXiv

  35. [43]

    Time-moe: Billion-scale time series foundation models with mixture of experts

    Xiaoming Shi, Shiyu Wang, Yuqi Nie, Dianqi Li, Zhou Ye, Qingsong Wen, and Ming Jin. Time-moe: Billion-scale time series foundation models with mixture of experts. InProceedings of the 13th International Conference on Learning Representations (ICLR 2025), May 2025. Spotlight

    2025

  36. [44]

    Well googled is half done: Multimodal forecasting of new fashion product sales with image-based google trends.Journal of Forecasting, 43(6):1982–1997, 2024

    Geri Skenderi, Christian Joppi, Matteo Denitto, and Marco Cristani. Well googled is half done: Multimodal forecasting of new fashion product sales with image-based google trends.Journal of Forecasting, 43(6):1982–1997, 2024

    1982

  37. [45]

    The multi-modal universe of fast-fashion: The visuelle 2.0 benchmark

    Geri Skenderi, Christian Joppi, Matteo Denitto, Berniero Scarpa, and Marco Cristani. The multi-modal universe of fast-fashion: The visuelle 2.0 benchmark. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 2241–2246, June 2022

    2022

  38. [46]

    TEST: Text prototype aligned embedding to activate LLM’s ability for time series.arXiv preprint arXiv:2308.08241, 2023

    Chenxi Sun, Hongyan Li, Yaliang Li, and Shenda Hong. TEST: Text prototype aligned embedding to activate LLM’s ability for time series.arXiv preprint arXiv:2308.08241, 2023

    2023 arXiv

  39. [47]

    Ptb-xl, a large publicly available electrocardiography dataset.Scientific Data, 7(1):154, 2020

    Patrick Wagner, Nils Strodthoff, Ralf-Dieter Bousseljot, David Kreiseler, Felix I Lunze, Wo- jciech Samek, and Tobias Schaeffter. Ptb-xl, a large publicly available electrocardiography dataset.Scientific Data, 7(1):154, 2020

    2020

  40. [48]

    Moirai: Scaling time series forecasting with foundation models

    Yuyang Wang, Bryan Lim, et al. Moirai: Scaling time series forecasting with foundation models. arXiv preprint arXiv:2403.14735, 2024

    2024 arXiv

  41. [49]

    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. InProceedings of the 41st International Conference on Machine Learning (ICML 2024), July 2024. Oral Presentation. 22

    2024

  42. [50]

    Gift-eval: A benchmark for general time series forecasting model evaluation.arXiv preprint arXiv:2410.10393, 2024

    Gerald Woo, Xu Liu, Taha Aksu, et al. Gift-eval: A benchmark for general time series forecasting model evaluation.arXiv preprint arXiv:2410.10393, 2024

    2024 arXiv

  43. [51]

    Dual-forecaster: A multimodal time series model integrating descriptive and predictive texts.arXiv preprint arXiv:2505.01135, 2025

    Wenfa Wu, Guanyu Zhang, Tan Zheng, Yi Wang, and Hongsheng Qi. Dual-forecaster: A multimodal time series model integrating descriptive and predictive texts.arXiv preprint arXiv:2505.01135, 2025

    2025 arXiv

  44. [52]

    Hanzhong Xia, Xiaoxia Chen, Zhen Wang, Xinyi Chen, and Fangyan Dong. A multi-modal deep-learning air quality prediction method based on multi-station time-series data and remote- sensing images: Case study of beijing and tianjin.Entropy, 26(1):91, 2024

    2024

  45. [53]

    Web traffic forecasting with cnn and fibonacci median

    Louis Yang and Chen-Hsi Huang. Web traffic forecasting with cnn and fibonacci median. https://github.com/louis925/kaggle-web-traffic-time-series-forecasting ,

  46. [54]

    Community trend prediction on heterogeneous graph in e-commerce

    Jiahao Yuan, Zhao Li, Pengcheng Zou, Xuan Gao, Jinwei Pan, Wendi Ji, and Xiaoling Wang. Community trend prediction on heterogeneous graph in e-commerce. InProceedings of the 15th ACM International Conference on Web Search and Data Mining (WSDM ’22), pages 1319–1327. ACM, February 2022

    2022

  47. [55]

    Are transformers effective for time series forecasting?arXiv preprint arXiv:2205.13504, 2022

    Ailing Zeng, Muxi Chen, Lei Zhang, and Qiang Xu. Are transformers effective for time series forecasting?arXiv preprint arXiv:2205.13504, 2022

    2022 arXiv

  48. [56]

    Moment: A foundation model for multivariate time series

    Han Zhang, Cheng Xiong, et al. Moment: A foundation model for multivariate time series. arXiv preprint arXiv:2405.02358, 2024

    2024

  49. [57]

    Deep time series models: A comprehensive survey and benchmark.arXiv preprint arXiv:2407.13278, 2024

    Qingsong Zhang, Haoxin Liu, Shangqing Xu, Zhiyuan Zhao, Lingkai Kong, et al. Deep time series models: A comprehensive survey and benchmark.arXiv preprint arXiv:2407.13278, 2024

    2024 arXiv

  50. [58]

    From news to forecast: Integrating event analysis in llm-based time series forecasting with reflection

    Junhua Zhao, Xinlei Wang, Jinjin Gu, and Hongming Zhang. From news to forecast: Integrating event analysis in llm-based time series forecasting with reflection. InAdvances in Neural Information Processing Systems, volume 37, pages 1–14, 2024

    2024

  51. [59]

    Lemma-rca: A large multi-modal multi-domain dataset for root cause analysis, 2024

    Lecheng Zheng, Zhengzhang Chen, Dongjie Wang, Chengyuan Deng, Reon Matsuoka, and Haifeng Chen. Lemma-rca: A large multi-modal multi-domain dataset for root cause analysis, 2024

    2024

  52. [60]

    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 (AAAI), pages 11106–11115, 2021

    2021

  53. [61]

    Ccwtf: Codebase for cold-start and varying-length time series forecasting

    Xin Zhou. Ccwtf: Codebase for cold-start and varying-length time series forecasting. https: //github.com/xinzzzhou/CCWTF, 2024. Accessed: 2025-05-12

    2024

  54. [62]

    Scalable transformer for high dimensional multivariate time series forecast- ing

    Xin Zhou, Weiqing Wang, Wray Buntine, Shilin Qu, Abishek Sriramulu, Weicong Tan, and Christoph Bergmeir. Scalable transformer for high dimensional multivariate time series forecast- ing. InProceedings of the 33rd ACM International Conference on Information and Knowledge Manage...

    2024

  55. [63]

    Unveiling the potential of text in high-dimensional time series forecasting

    Xin Zhou, Weiqing Wang, Shilin Qu, Zhang Zhiqiang, and Christoph Bergmeir. Unveiling the potential of text in high-dimensional time series forecasting. InProceedings of the NeurIPS 2024 Workshop on Time Series in the Age of Large Models (TSALM), 2024. 23

    2024

  56. [2017]

    Kaggle Competition Solution (Ranked 11th/1095)

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.