Pith. sign in

REVIEW 4 major objections 6 minor 138 references

Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectives

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read TimesCLIP renders each time series as both a colorized plot and a token sequence, aligns the two views with contrastive learning, and reports consistent forecasting gains over unimodal and LLM-based baselines.

desk verdict TimesCLIP has a genuinely novel idea and some real gains, but the paper's headline claim of consistent SOTA is not supported by its own ablations. read the letter →

arxiv 2506.24124 v2 pith:M6NOLTIB submitted 2025-06-30 cs.LG cs.CV

Sixun Dong , Wei Fan , Teresa Wu , Yanjie Fu This is my paper
classification cs.LGcs.CV
keywords timeseriesforecastingmultimodalcontrastivelearningvision-languagealignmentCLIPvariateselectionpatchtokenizationlong-termshort-term
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

This paper tries to show that a time series carries more usable signal when a model can both see it and read it. TimesCLIP renders each variate as a color-coded line plot (the visual view) and also slices it into patch tokens that a pretrained text encoder consumes (the textual view), then uses a bidirectional InfoNCE contrastive loss to pull the two views into a shared space. The authors argue that this multimodal alignment captures semantic and perceptual patterns that purely numerical models miss, and that a learned class token from the aligned space can pick out the most informative variates for prediction. On fifteen short-term and six long-term benchmarks, they report that TimesCLIP beats strong unimodal baselines and LLM-based methods, including best results on the M4 weighted average and on all six long-term datasets.

What carries the argument

The load-bearing mechanism is the contrastive alignment of two synthetic views of the same series: a colorized line plot and a patchified token sequence, with a learnable [class] token on the language side serving as the alignment anchor. A bidirectional InfoNCE loss $\mathcal{L}_{\text{align}}$ pulls the visual class token and the textual class token of the same series together in a shared CLIP-style space, while the frozen pretrained encoders supply the perceptual and linguistic inductive biases. The aligned class token is then reused as the query in a cross-attention layer that selects the most informative variates, and the resulting feature is fused into the generator by replacing the last token of the language representation.

What would settle it

Run TimesCLIP on Traffic and ECL with the full vision branch and contrastive loss (using gradient checkpointing or a smaller visual encoder to fit memory) and compare with the reported numbers; if the full model does not match or beat the unimodal baselines, the claim that multimodal alignment drives the gains fails on exactly the datasets where the mechanism was omitted.

Watch

Extended reading notes

Core claim

The central discovery, on the authors' terms, is that multimodal contrastive alignment constructed entirely from numerical data, with no natural language and no real photographs, produces representations that forecast better than either modality alone or than standard numerical encoders. The vision branch normalizes each variate within a window, plots it in a fixed color, and passes it through a frozen CLIP vision encoder; the language branch layer-normalizes, patchifies, and sends the tokens through a learned tokenizer into the CLIP text encoder with a learnable [class] token. The two views are aligned with a bidirectional InfoNCE loss computed over variate-level class tokens, and a cross-attention variate selection module uses the aligned class token as a query to emphasize relevant variables before a linear head generates the forecast. The authors take the consistent wins on M4, EPF, PEMS, illness, and the six long-term benchmarks as evidence that aligning a visual and a textual perspective of the same series extracts complementary information that unimodal models leave unused.

Load-bearing premise

The strong long-term results on Traffic and ECL were produced by a modified TimesCLIP that discards the vision branch and the contrastive alignment loss to fit GPU memory, so the paper's central claim assumes that this modified model still tests the multimodal mechanism; if it does not, two of the six state-of-the-art long-term results are not evidence for the method the paper proposes.

Editorial extensions

If this is right

  • If the central claim holds, time series forecasting no longer needs to choose between numerical, visual, and textual representations: a contrastively aligned combination of the latter two is the stronger input.
  • Pretrained vision-language backbones such as CLIP can be reused for time series without prompting or natural-language text, requiring only a learned tokenizer and projection rather than a per-dataset architecture.
  • The variate selection module means the model can highlight which series matter for a multivariate prediction, which is directly useful for high-dimensional settings such as traffic or energy grids.
  • Ablations show that the alignment loss, colorization, and variate selection each contribute, so the reported performance depends on the full multimodal mechanism rather than on any single component.
  • Because CLIP-style multimodal pretraining helps more than text-only pretraining, the advantage appears to live in the multimodal space itself, not merely in having a language encoder.

Reading between the lines

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

  • If the alignment loss is the true driver, a straightforward test is to swap in a randomly initialized vision-text backbone while keeping everything else fixed: the gains should shrink if pretrained multimodal alignment matters, and should not if the improvement comes from extra parameters or from the image input alone.
  • The appendix drops the vision branch and contrastive loss on Traffic and ECL for GPU memory reasons, so the state-of-the-art claim on those two datasets is currently supported only by the text side of the model; re-running with an efficient vision branch would place the paper's central mechanism on firmer ground.
  • Colorizing each variate effectively encodes channel identity into the image; a similar one-hot channel embedding on the language side might reproduce part of the gain, which would clarify whether the benefit is genuinely visual or just a better way to mark which series is which.
  • The aligned class token is a general-purpose representation, so extending TimesCLIP to time series classification, anomaly detection, or imputation would test whether the multimodal space transfers beyond the forecasting objective it was trained on.
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. This paper proposes TimesCLIP, a forecasting model that builds two representations from a multivariate time series: a visual view obtained by plotting each normalized variate as a colorized line image and encoding it with a frozen CLIP vision encoder, and a textual/token view obtained by patchifying the series and passing the patches through a learnable tokenizer into a fine-tuned CLIP text encoder. The two views are aligned with a bidirectional InfoNCE contrastive loss, and the language-side class token is used in a cross-attention variate-selection module before a linear head produces the forecast. The authors report short-term results on M4, PEMS, illness, and EPF datasets and long-term results on Exchange, Traffic, Weather, ETTm1, ETTm2, and Solar-Energy, claiming that the approach consistently outperforms strong unimodal and LLM-based baselines and achieves the best performance on six long-term datasets.

Significance. If the reported gains were attributable to the proposed multimodal contrastive alignment, the paper would be a useful contribution: it demonstrates a way to repurpose frozen CLIP backbones for time series forecasting, it includes a useful ablation on Exchange (Table 5) showing that colorization and the contrastive loss matter, and it provides a backbone comparison and a FLOPs/parameter analysis. The paper is also transparent about resource constraints in Appendix G. However, the central claim is weakened in two ways. First, both modalities are deterministic functions of the same numerical input, so the contrastive alignment is a within-sample consistency constraint rather than an independent cross-modal grounding. Second, and more importantly, two of the six long-term datasets are evaluated with a model variant that omits the proposed mechanism entirely, and on two other datasets the reported numbers contradict the 'best on six datasets' claim. The Exchange ablation is a real effect, but one dataset cannot carry the 'consistently outperforms' claim made in the abstract.

major comments (4)
  1. [Section 4.3, Table 4, Appendix E.2, Appendix G] The long-term results for Traffic and ECL are attributed to TimesCLIP, but Appendix E.2 states that for these datasets the model was modified to embed numerical time series with the tokenizer only, to exclude the multimodal contrastive learning loss, and to rely solely on the pretrained multimodal language model; Appendix G repeats that the vision module was omitted to conserve GPU resources. Because the proposed mechanism consists precisely of the vision branch and the contrastive loss L_align, the Traffic and ECL rows do not test the method described in Sections 3.2 and 3.3, and they cannot be used to support the abstract's claim that multimodal alignment consistently outperforms baselines.
  2. [Table 4, Section 4.3] Table 4 shows that on ETTm1 the average MSE of TimesCLIP is 0.343, worse than Time-LLM's 0.318, and on ETTm2 it is 0.209, worse than GPT4TS's 0.193; on Solar-Energy the average MSE is tied with iTransformer and PatchTST at 0.224. Consequently, the statement in Section 4.3 that TimesCLIP 'achieves the best performance on six datasets' is not supported by the table's own numbers, and the claim needs to be narrowed or the experiments need to be corrected.
  3. [Section 4.3, Table 4] Section 4.3 says the long-term evaluation covers 'six standard long-term forecasting datasets: Electricity (ECL), Exchange, Traffic, Weather, Illness, ETTm1, and ETTm2,' which is a list of seven datasets, while Table 4 reports Exchange, Traffic, Weather, ETTm1, ETTm2, and Solar-Energy and contains no ECL or Illness entries. In addition, the table caption refers to an average over 'all four prediction lengths' although only 96 and 192 are evaluated. These inconsistencies make the 'best on six datasets' claim ambiguous and need to be resolved before the evaluation can be assessed.
  4. [Sections 3.2, 3.3, 5.1] Because the visual and textual views are both deterministic transformations of the same numerical input, the contrastive loss L_align in Eq. (11) enforces consistency between two views of the same signal rather than grounding the representation in an independent modality. The improvement in Table 5 from 0.389 to 0.335 MSE on Exchange could therefore come from the additional capacity of the frozen ViT or from a regularizing consistency objective rather than from cross-modal alignment. To substantiate the claim that multimodal alignment is the source of the gains, the authors should add capacity-matched unimodal baselines or a control with randomized alignment targets.
minor comments (6)
  1. [Section 4.2] The count of short-term benchmarks is unclear: M4 contributes six frequencies, PEMS contributes four, EPF contributes five, and illness is reported separately, which totals sixteen rather than the claimed fifteen.
  2. [Equation (9)] Equation (9) uses the notation X^{img}_{cls,i} but the vision feature was defined earlier as CLS^{img}_i = Proj_img(f_i); please align the notation between Eq. (9) and the text.
  3. [Table 2] Table 2 lists 'PEMS03' twice; the third row should presumably be PEMS08 based on Table 10.
  4. [Section 4.3, Appendix E.2] Section 4.3 says the paper intentionally omits prediction lengths 336 and 720, but Appendix E.2's Table 11 lists them as evaluated prediction lengths for Traffic and ECL; please reconcile these statements.
  5. [Section 4.1, Section 3.2] Section 4.1 states that CLIP's original token embedding is replaced by a new embedding trained from scratch, while Section 3.2 describes the tokenizer as a randomly initialized learnable linear layer; the relationship between these two descriptions should be clarified.
  6. [References] The bibliography contains several duplicate entries (e.g., [41]/[42], [92]/[93], [115]/[116], [137]/[138], [139]/[140]) and a number of works that appear unrelated to the paper's subject matter; please clean up the reference list.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: forecasting is evaluated on held-out test sets and no load-bearing step reduces to its inputs; the Traffic/ECL appendix disclosure is a correctness concern, not circularity.

full rationale

The paper's derivation chain is not circular. The two 'modalities' (colorized images and tokenized patches) are both constructed from the same numerical input, and the InfoNCE loss Lalign is indeed a within-sample consistency regularizer; but that is a modeling choice, not a circular prediction. Forecasting is performed by a generator trained with Lgen on training splits and evaluated on held-out test sets of M4, PEMS, EPF, and long-term benchmarks, so the headline results are not forced by construction. The variate-selection cross-attention (Eq. 13) uses the language CLS token as query and variate representations as key/value; it is an attention mechanism, not a definitional reuse of the target. Citations to CLIP, CoCa, PatchTST, and Time-LLM are external baselines and foundations; the paper's self-citations in Appendix G are illustrative downstream applications and are not load-bearing. The paper's own Appendix E.2 discloses that the Traffic and ECL long-term results were produced by a variant that omits the vision branch and excludes Lalign, and Table 4 does not show TimesCLIP beating Time-LLM or GPT4TS on ETTm1/ETTm2; these are correctness and evidence-quality concerns about how strongly the central multimodal claim is supported, not circularity. A claim can be overstated without being circular.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central method relies on several ad hoc modeling choices: the visualization pipeline (color, normalization window, figure rendering) is a free parameter that is not specified, and the use of a frozen image encoder plus a fine-tuned text encoder assumes that pretrained vision-language models transfer to time series. No new physical entities are introduced.

free parameters (2)
  • λ2 (contrastive loss weight) = 0.1 default; 0.5, 0.1, 0.1, 0.9 for Exchange by horizon
    The weight of Lalign is set per dataset and per prediction length in Appendix E.2, which resembles tuning to each benchmark rather than a fixed choice.
  • Lwindow (normalization window length) = Not reported
    The visualization preprocessing normalizes each variable in a fixed window of size Lwindow, but the paper never specifies this value, leaving the image generation step under-determined.
assumptions (3)
  • ad hoc to paper Colorized line plots of normalized time series, when encoded by a frozen ViT, preserve forecasting-relevant structure.
    The entire vision branch depends on this assumption; the paper provides an ablation on Exchange but no analysis of which visual features matter.
  • domain assumption A random-initialized linear tokenizer plus fine-tuning of the CLIP text encoder can adapt numerical time series patches into a meaningful text-like embedding space.
    Section 3.2 uses a learned tokenizer with a pretrained text encoder, but the paper does not validate that the resulting embeddings are semantically coherent.
  • domain assumption Instance-level contrastive pairs (same sample's image and text views) form a useful training signal for forecasting.
    The contrastive loss treats the two synthetic views of the same series as a positive pair; this is a standard self-supervised assumption, but its value for forecasting is only empirically tested on two datasets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectives." pith.science (2026). https://pith.science/paper/M6NOLTIB

@misc{pith2026250624124,
  author       = {Pith},
  title        = {Pith review of: Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectives},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M6NOLTIB}},
  note         = {Machine review of arXiv:2506.24124}
}
read the original abstract

Time series forecasting traditionally relies on unimodal numerical inputs, which often struggle to capture high-level semantic patterns due to their dense and unstructured nature. While recent approaches have explored representing time series as text using large language models (LLMs), these methods remain limited by the discrete nature of token sequences and lack the perceptual intuition humans typically apply, such as interpreting visual patterns. In this paper, we propose a multimodal contrastive learning framework that transforms raw time series into structured visual and textual perspectives. Rather than using natural language or real-world images, we construct both modalities directly from numerical sequences. We then align these views in a shared semantic space via contrastive learning, enabling the model to capture richer and more complementary representations. Furthermore, we introduce a variate selection module that leverages the aligned representations to identify the most informative variables for multivariate forecasting. Extensive experiments on fifteen short-term and six long-term forecasting benchmarks demonstrate that our approach consistently outperforms strong unimodal and cross-modal baselines, highlighting the effectiveness of multimodal alignment in enhancing time series forecasting. Code is available at: https://github.com/Ironieser/TimesCLIP.

Figures

Figures reproduced from arXiv: 2506.24124 by the authors.

Figure 1
Figure 1. Comparison of frameworks and modeling strategies. Overview of modeling paradigms for [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of the TimesCLIP framework. TimesCLIP converts numerical multivariate time [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The visualization of Visualization Preprocess. Sample data is from Weather dataset. [PITH_FULL_IMAGE:figures/full_fig_p021_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

138 extracted references · 33 canonical work pages

  1. [1]

    Maddix, Hao Wang, Michael W

    Abdul Fatir Ansari, Lorenzo Stella, Caner Turkmen, Xiyuan Zhang, Pedro Mercado, Huibin Shen, Oleksandr Shchur, Syama Sundar Rangapuram, Sebastian Pineda Arango, Shubham Kapoor, Jasper Zschiegner, Danielle C. Maddix, Hao Wang, Michael W. Mahoney, Kari Torkkola, Andrew Gordon Wilson, Michael Bohlke-Schneider, and Yuyang Wang. Chronos: Learning the language ...

    arXiv 2024

  2. [2]

    Layer normalization

    Jimmy Lei Ba. Layer normalization. arXiv preprint arXiv:1607.06450, 2016

    arXiv 2016

  3. [3]

    Privacy preserving generative feature transformation

    Haoyue Bai, Wangyang Ying, Nanxu Gong, Xinyuan Wang, Hao Liu, and Yanjie Fu. Privacy preserving generative feature transformation

  4. [4]

    Gorec: a generative cold-start recommendation framework

    Haoyue Bai, Min Hou, Le Wu, Yonghui Yang, Kun Zhang, Richang Hong, and Meng Wang. Gorec: a generative cold-start recommendation framework. In Proceedings of the 31st ACM international conference on multimedia, pages 1004–1012, 2023

    2023

  5. [5]

    Multimodality invariant learning for multimedia-based new item recommendation

    Haoyue Bai, Le Wu, Min Hou, Miaomiao Cai, Zhuangzhuang He, Yuyang Zhou, Richang Hong, and Meng Wang. Multimodality invariant learning for multimedia-based new item recommendation. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 677–686, 2024

    2024

  6. [6]

    Brownian bridge augmented surrogate simulation and injection planning for geological co _2 storage

    Haoyue Bai, Guodong Chen, Wangyang Ying, Xinyuan Wang, Nanxu Gong, Sixun Dong, Giulia Pedrielli, Haoyu Wang, Haifeng Chen, and Yanjie Fu. Brownian bridge augmented surrogate simulation and injection planning for geological co _2 storage. arXiv preprint arXiv:2505.18204, 2025

    arXiv 2025

  7. [7]

    Deep learning and time series-to-image encoding for finan- cial forecasting

    Silvio Barra, Salvatore Mario Carta, Andrea Corriga, Alessandro Sebastian Podda, and Diego Reforgiato Recupero. Deep learning and time series-to-image encoding for finan- cial forecasting. IEEE/CAA Journal of Automatica Sinica, 7(3):683–692, 2020

    2020

  8. [8]

    Fundamental limitations of foundational forecasting models: The need for multimodality and rigorous evaluation

    Christoph Bergmeir. Fundamental limitations of foundational forecasting models: The need for multimodality and rigorous evaluation. Invited Talk at the NeurIPS 2024 Workshop on Time Series in the Age of Large Models, December 2024. https://neurips.cc/virtual/ 2024/108471

    2024

Show all 138 references
  1. [9]

    Control charts in financial applications: An overview

    Konstantinos Bisiotis, Stelios Psarakis, and Athanasios N Yannacopoulos. Control charts in financial applications: An overview. Quality and Reliability Engineering International, 38(3): 1441–1462, 2022

    2022

  2. [10]

    Language models are few-shot learners

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhari- wal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877– 1901, 2020

    1901

  3. [11]

    Time series forecasting for healthcare diagnosis and prognostics with the focus on cardiovascular diseases

    C Bui, N Pham, A V o, A Tran, A Nguyen, and T Le. Time series forecasting for healthcare diagnosis and prognostics with the focus on cardiovascular diseases. In 6th International Conference on the Development of Biomedical Engineering in Vietnam (BME6) 6 , pages 809–818. Sprin...

    2018

  4. [12]

    From orthogonality to dependency: Learning disentangled representation for multi-modal time-series sensing signals

    Ruichu Cai, Zhifang Jiang, Zijian Li, Weilin Chen, Xuexin Chen, Zhifeng Hao, Yifan Shen, Guangyi Chen, and Kun Zhang. From orthogonality to dependency: Learning disentangled representation for multi-modal time-series sensing signals. arXiv preprint arXiv:2405.16083, 2024

    2024 arXiv

  5. [13]

    Lightts: Lightweight time series classification with adaptive ensemble distillation

    David Campos, Miao Zhang, Bin Yang, Tung Kieu, Chenjuan Guo, and Christian S Jensen. Lightts: Lightweight time series classification with adaptive ensemble distillation. Proceedings of the ACM on Management of Data, 1(2):1–27, 2023

    2023

  6. [14]

    Locvtp: Video-text pre-training for temporal localization

    Meng Cao, Tianyu Yang, Junwu Weng, Can Zhang, Jue Wang, and Yuexian Zou. Locvtp: Video-text pre-training for temporal localization. arXiv preprint arXiv:2207.10362, 2022. 10

    2022 arXiv

  7. [15]

    Nhits: Neural hierarchical interpolation for time series forecasting

    Cristian Challu, Kin G Olivares, Boris N Oreshkin, Federico Garza Ramirez, Max Mergenthaler Canseco, and Artur Dubrawski. Nhits: Neural hierarchical interpolation for time series forecasting. In Proceedings of the AAAI conference on artificial intelligence, volume 37, pages 69...

    2023

  8. [16]

    Multi- model approach for stock price prediction and trading recommendations

    Zhenrui Chen, Zhibo Dai, Huiyan Xing, Junyu Chen, Menghao Huo, and Kuan Lu. Multi- model approach for stock price prediction and trading recommendations. Preprints, May

  9. [17]

    Financial time series forecasting with multi-modality graph neural network

    Dawei Cheng, Fangzhou Yang, Sheng Xiang, and Jin Liu. Financial time series forecasting with multi-modality graph neural network. Pattern Recognition, 121:108218, 2022

    2022

  10. [18]

    Long-term forecasting with tide: Time-series dense encoder

    Abhimanyu Das, Weihao Kong, Andrew Leach, Shaan Mathur, Rajat Sen, and Rose Yu. Long-term forecasting with tide: Time-series dense encoder. arXiv preprint arXiv:2304.08424, 2023

    2023 arXiv

  11. [19]

    Timesiam: A pre-training framework for siamese time-series modeling

    Jiaxiang Dong, Haixu Wu, Yuxuan Wang, Yunzhong Qiu, Li Zhang, Jianmin Wang, and Mingsheng Long. Timesiam: A pre-training framework for siamese time-series modeling. arXiv preprint arXiv:2402.02475, 2024

    2024 arXiv

  12. [20]

    Weakly supervised video representation learning with unaligned text for sequential videos

    Sixun Dong, Huazhang Hu, Dongze Lian, Weixin Luo, Yicheng Qian, and Shenghua Gao. Weakly supervised video representation learning with unaligned text for sequential videos. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2437–2447, 2023

    2023

  13. [21]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv...

    2010 arXiv

  14. [22]

    Lanistr: Multimodal learning from structured and unstructured data

    Sayna Ebrahimi, Sercan O Arik, Yihe Dong, and Tomas Pfister. Lanistr: Multimodal learning from structured and unstructured data. arXiv preprint arXiv:2305.16556, 2023

    2023 arXiv

  15. [23]

    Unsupervised scalable repre- sentation learning for multivariate time series

    Jean-Yves Franceschi, Aymeric Dieuleveut, and Martin Jaggi. Unsupervised scalable repre- sentation learning for multivariate time series. Advances in neural information processing systems, 32, 2019

    2019

  16. [24]

    Agentic feature augmentation: Unifying selection and generation with teaming, planning, and memories

    Nanxu Gong, Sixun Dong, Haoyue Bai, Xinyuan Wang, Wangyang Ying, and Yanjie Fu. Agentic feature augmentation: Unifying selection and generation with teaming, planning, and memories. arXiv preprint arXiv:2505.15076, 2025

    2025 arXiv

  17. [25]

    Sculpting features from noise: Reward-guided hierarchical diffusion for task-optimal feature transformation

    Nanxu Gong, Zijun Li, Sixun Dong, Haoyue Bai, Wangyang Ying, Xinyuan Wang, and Yanjie Fu. Sculpting features from noise: Reward-guided hierarchical diffusion for task-optimal feature transformation. arXiv preprint arXiv:2505.15152, 2025

    2025 arXiv

  18. [26]

    Evolutionary large language model for automated feature transformation

    Nanxu Gong, Chandan K Reddy, Wangyang Ying, Haifeng Chen, and Yanjie Fu. Evolutionary large language model for automated feature transformation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 16844–16852, 2025

    2025

  19. [27]

    Unsupervised feature transformation via in-context generation, generator-critic llm agents, and duet-play teaming

    Nanxu Gong, Xinyuan Wang, Wangyang Ying, Haoyue Bai, Sixun Dong, Haifeng Chen, and Yanjie Fu. Unsupervised feature transformation via in-context generation, generator-critic llm agents, and duet-play teaming. arXiv preprint arXiv:2504.21304, 2025

    2025 arXiv

  20. [28]

    Neuro-symbolic embedding for short and effective feature selection via autoregressive generation

    Nanxu Gong, Wangyang Ying, Dongjie Wang, and Yanjie Fu. Neuro-symbolic embedding for short and effective feature selection via autoregressive generation. ACM Transactions on Intelligent Systems and Technology, 16(2):1–21, 2025

    2025

  21. [29]

    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.. 2024. arXiv preprint arXiv:2402.03885, 2024. 11

    2024 arXiv

  22. [30]

    Large language models are zero-shot time series forecasters

    Nate Gruver, Marc Finzi, Shikai Qiu, and Andrew G Wilson. Large language models are zero-shot time series forecasters. Advances in Neural Information Processing Systems, 36, 2024

    2024

  23. [31]

    Efficiently modeling long sequences with structured state spaces

    Albert Gu, Karan Goel, and Christopher Ré. Efficiently modeling long sequences with structured state spaces. arXiv preprint arXiv:2111.00396, 2021

    2021 arXiv

  24. [32]

    Audioclip: Extending clip to image, text and audio

    Andrey Guzhov, Federico Raue, Jörn Hees, and Andreas Dengel. Audioclip: Extending clip to image, text and audio. In ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 976–980. IEEE, 2022

    2022

  25. [33]

    Temporal alignment networks for long- term video

    Tengda Han, Weidi Xie, and Andrew Zisserman. Temporal alignment networks for long- term video. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2906–2916, 2022

    2022

  26. [34]

    Parameter-efficient fine-tuning for large models: A comprehensive survey

    Zeyu Han, Chao Gao, Jinyang Liu, Jeff Zhang, and Sai Qian Zhang. Parameter-efficient fine-tuning for large models: A comprehensive survey. arXiv preprint arXiv:2403.14608, 2024

    2024 arXiv

  27. [35]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016

    2016

  28. [36]

    Double correction framework for denoising recommendation

    Zhuangzhuang He, Yifan Wang, Yonghui Yang, Peijie Sun, Le Wu, Haoyue Bai, Jinqi Gong, Richang Hong, and Min Zhang. Double correction framework for denoising recommendation. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 1062–1072, 2024

    2024

  29. [37]

    Long short-term memory

    S Hochreiter. Long short-term memory. Neural Computation MIT-Press, 1997

    1997

  30. [38]

    Transrac: Encoding multi-scale temporal correlation with transformers for repetitive action counting

    Huazhang Hu, Sixun Dong, Yiqun Zhao, Dongze Lian, Zhengxin Li, and Shenghua Gao. Transrac: Encoding multi-scale temporal correlation with transformers for repetitive action counting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19...

    2022

  31. [39]

    Ct-patchtst: Channel-time patch time-series transformer for long-term renewable energy forecasting

    Menghao Huo, Kuan Lu, Yuxiao Li, and Qiang Zhu. Ct-patchtst: Channel-time patch time-series transformer for long-term renewable energy forecasting. arXiv preprint arXiv:2501.08620, 2025. URL https://arxiv.org/abs/2501.08620

    2025

  32. [40]

    Gpt4mts: Prompt-based large language model for multimodal time-series forecasting

    Furong Jia, Kevin Wang, Yixiang Zheng, Defu Cao, and Yan Liu. Gpt4mts: Prompt-based large language model for multimodal time-series forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 23343–23351, 2024

    2024

  33. [42]

    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, 2024. URL https://arxiv.org/ abs/2310.01728

    2024 arXiv

  34. [43]

    Position: What can large language models tell us about time series analysis

    Ming Jin, Yifan Zhang, Wei Chen, Kexin Zhang, Yuxuan Liang, Bin Yang, Jindong Wang, Shirui Pan, and Qingsong Wen. Position: What can large language models tell us about time series analysis. In Forty-first International Conference on Machine Learning, 2024

    2024

  35. [44]

    Ai in healthcare: time-series forecasting using statistical, neural, and ensemble architectures

    Shruti Kaushik, Abhinav Choudhury, Pankaj Kumar Sheron, Nataraj Dasgupta, Sayee Natara- jan, Larry A Pickett, and Varun Dutt. Ai in healthcare: time-series forecasting using statistical, neural, and ensemble architectures. Frontiers in big data, 3:4, 2020

    2020

  36. [45]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin Ming-Wei Chang Kenton and Lee Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of naacL-HLT, volume 1, page 2. Minneapolis, Minnesota, 2019. 12

    2019

  37. [46]

    Reformer: The efficient transformer

    Nikita Kitaev, Łukasz Kaiser, and Anselm Levskaya. Reformer: The efficient transformer. arXiv preprint arXiv:2001.04451, 2020

    2001 arXiv

  38. [47]

    Lite: Modeling environmental ecosystems with multimodal large language models

    Haoran Li, Junqi Liu, Zexian Wang, Shiyuan Luo, Xiaowei Jia, and Huaxiu Yao. Lite: Modeling environmental ecosystems with multimodal large language models. arXiv preprint arXiv:2404.01165, 2024

    2024 arXiv

  39. [48]

    Sehf: A summary- enhanced hierarchical framework for financial report sentiment analysis

    Haozhou Li, Qinke Peng, Xinyuan Wang, Xu Mou, and Yonghao Wang. Sehf: A summary- enhanced hierarchical framework for financial report sentiment analysis. IEEE Transactions on Computational Social Systems, 11(3):4087–4101, 2023

    2023

  40. [49]

    Sade: A speaker- aware dual encoding model based on diagbert for medical triage and pre-diagnosis

    Haozhou Li, Xinyuan Wang, Hongkai Du, Wentong Sun, and Qinke Peng. Sade: A speaker- aware dual encoding model based on diagbert for medical triage and pre-diagnosis. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Process- ing (ICASSP), pages ...

    2024

  41. [50]

    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, pages 402–415. PMLR, 2024

    2024

  42. [51]

    Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models. In International conference on machine learning, pages 19730–19742. PMLR, 2023

    2023

  43. [52]

    Videochat: Chat-centric video understanding

    KunChang Li, Yinan He, Yi Wang, Yizhuo Li, Wenhai Wang, Ping Luo, Yali Wang, Limin Wang, and Yu Qiao. Videochat: Chat-centric video understanding. arXiv preprint arXiv:2305.06355, 2023

    2023 arXiv

  44. [53]

    Clip-event: Connecting text and images with event structures

    Manling Li, Ruochen Xu, Shuohang Wang, Luowei Zhou, Xudong Lin, Chenguang Zhu, Michael Zeng, Heng Ji, and Shih-Fu Chang. Clip-event: Connecting text and images with event structures. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16...

    2022

  45. [54]

    Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting

    Shiyang Li, Xiaoyong Jin, Yao Xuan, Xiyou Zhou, Wenhu Chen, Yu-Xiang Wang, and Xifeng Yan. Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. Advances in neural information processing systems, 32, 2019

    2019

  46. [55]

    Deep learning models for time series forecasting: a review

    Wenxiang Li and KL Eddie Law. Deep learning models for time series forecasting: a review. IEEE Access, 2024

    2024

  47. [56]

    Forecasting with time series imaging

    Xixi Li, Yanfei Kang, and Feng Li. Forecasting with time series imaging. Expert Systems with Applications, 160:113680, 2020

    2020

  48. [57]

    Time series as images: Vision transformer for irregularly sampled time series

    Zekun Li, Shiyang Li, and Xifeng Yan. Time series as images: Vision transformer for irregularly sampled time series. Advances in Neural Information Processing Systems , 36, 2024

    2024

  49. [58]

    Revisiting long-term time series forecasting: An investigation on linear mapping

    Zhe Li, Shiyi Qi, Yiduo Li, and Zenglin Xu. Revisiting long-term time series forecasting: An investigation on linear mapping. arXiv preprint arXiv:2305.10721, 2023

    2023 arXiv

  50. [59]

    Segrnn: Segment recurrent neural network for long-term time series forecasting.arXiv preprint arXiv:2308.11200, 2023

    Shengsheng Lin, Weiwei Lin, Wentai Wu, Feiyu Zhao, Ruichao Mo, and Haotong Zhang. Segrnn: Segment recurrent neural network for long-term time series forecasting.arXiv preprint arXiv:2308.11200, 2023

    2023 arXiv

  51. [60]

    Pth and the regulation of mesenchymal cells within the bone marrow niche

    Hanghang Liu, Linyi Liu, and Clifford J Rosen. Pth and the regulation of mesenchymal cells within the bone marrow niche. Cells, 13(5):406, 2024

    2024

  52. [61]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. Advances in neural information processing systems, 36, 2024

    2024

  53. [62]

    Lstprompt: Large language models as zero-shot time series forecasters by long-short-term prompting

    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. arXiv preprint arXiv:2402.16132, 2024. 13

    2024 arXiv

  54. [63]

    Edta enhances stromal cell–derived factor 1α–induced migration of dental pulp cells by up-regulating chemokine receptor 4 expression

    Linyi Liu, Sha Leng, Junli Yue, Qian Lu, Weizhe Xu, Xiaowei Yi, Dingming Huang, and Lan Zhang. Edta enhances stromal cell–derived factor 1α–induced migration of dental pulp cells by up-regulating chemokine receptor 4 expression. Journal of Endodontics, 45(5):599–605, 2019

    2019

  55. [64]

    Calorie restriction in mice impairs cortical but not trabecular peak bone mass by suppressing bone remodeling

    Linyi Liu, Phuong T Le, J Patrizia Stohn, Hanghang Liu, Wangyang Ying, Roland Baron, and Clifford J Rosen. Calorie restriction in mice impairs cortical but not trabecular peak bone mass by suppressing bone remodeling. Journal of Bone and Mineral Research, 39(8):1188–1199, 2024

    2024

  56. [65]

    Scinet: Time series modeling and forecasting with sample convolution and interaction

    Minhao Liu, Ailing Zeng, Muxi Chen, Zhijian Xu, Qiuxia Lai, Lingna Ma, and Qiang Xu. Scinet: Time series modeling and forecasting with sample convolution and interaction. Advances in Neural Information Processing Systems, 35:5816–5828, 2022

    2022

  57. [66]

    Focal: Contrastive learning for multimodal time- series sensing signals in factorized orthogonal latent space

    Shengzhong Liu, Tomoyoshi Kimura, Dongxin Liu, Ruijie Wang, Jinyang Li, Suhas Diggavi, Mani Srivastava, and Tarek Abdelzaher. Focal: Contrastive learning for multimodal time- series sensing signals in factorized orthogonal latent space. Advances in Neural Information Processin...

    2024

  58. [67]

    Pyraformer: Low-complexity pyramidal attention for long-range time series modeling and forecasting

    Shizhan Liu, Hang Yu, Cong Liao, Jianguo Li, Weiyao Lin, Alex X Liu, and Schahram Dustdar. Pyraformer: Low-complexity pyramidal attention for long-range time series modeling and forecasting. In # PLACEHOLDER_PARENT_METADATA_VALUE#, 2022

    2022

  59. [68]

    Unitime: A language-empowered unified model for cross-domain time series forecasting

    Xu Liu, Junfeng Hu, Yuan Li, Shizhe Diao, Yuxuan Liang, Bryan Hooi, and Roger Zim- mermann. Unitime: A language-empowered unified model for cross-domain time series forecasting. In Proceedings of the ACM on Web Conference 2024, pages 4095–4106, 2024

    2024

  60. [69]

    Non-stationary transformers: Ex- ploring the stationarity in time series forecasting

    Yong Liu, Haixu Wu, Jianmin Wang, and Mingsheng Long. Non-stationary transformers: Ex- ploring the stationarity in time series forecasting. Advances in Neural Information Processing Systems, 35:9881–9893, 2022

    2022

  61. [70]

    itransformer: Inverted transformers are effective for time series forecasting, 2024

    Yong Liu, Tengge Hu, Haoran Zhang, Haixu Wu, Shiyu Wang, Lintao Ma, and Mingsheng Long. itransformer: Inverted transformers are effective for time series forecasting, 2024. URL https://arxiv.org/abs/2310.06625

    2024 arXiv

  62. [71]

    Auto- times: Autoregressive time series forecasters via large language models

    Yong Liu, Guo Qin, Xiangdong Huang, Jianmin Wang, and Mingsheng Long. Auto- times: Autoregressive time series forecasters via large language models. arXiv preprint arXiv:2402.02370, 2024

    2024 arXiv

  63. [72]

    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

  64. [73]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 10012–10022, 2021

    2021

  65. [74]

    A cnn-bilstm-am method for stock price prediction

    Wenjie Lu, Jiazheng Li, Jingyang Wang, and Lele Qin. A cnn-bilstm-am method for stock price prediction. Neural Computing and Applications, 33(10):4741–4753, 2021

    2021

  66. [75]

    Howto100m: Learning a text-video embedding by watching hundred million narrated video clips

    Antoine Miech, Dimitri Zhukov, Jean-Baptiste Alayrac, Makarand Tapaswi, Ivan Laptev, and Josef Sivic. Howto100m: Learning a text-video embedding by watching hundred million narrated video clips. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages ...

    2019

  67. [76]

    Expanding language-image pretrained models for general video recognition

    Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, and Haibin Ling. Expanding language-image pretrained models for general video recognition. In European Conference on Computer Vision, pages 1–18. Springer, 2022

    2022

  68. [77]

    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. arXiv preprint arXiv:2211.14730, 2022. 14

    2022 arXiv

  69. [78]

    Representation learning with contrastive predictive coding

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748, 2018

    2018 arXiv

  70. [79]

    N-beats: Neural basis expansion analysis for interpretable time series forecasting

    Boris N Oreshkin, Dmitri Carpov, Nicolas Chapados, and Yoshua Bengio. N-beats: Neural basis expansion analysis for interpretable time series forecasting. arXiv preprint arXiv:1905.10437, 2019

    1905 arXiv

  71. [80]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32, 2019

    2019

  72. [81]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning, pa...

    2021

  73. [82]

    Exploring the limits of transfer learning with a unified text-to-text transformer

    Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research, 21(140):1–67, 2020

    2020

  74. [83]

    Automatic diagnosis of the 12-lead ecg using a deep neural network

    Antônio H Ribeiro, Manoel Horta Ribeiro, Gabriela MM Paixão, Derick M Oliveira, Paulo R Gomes, Jéssica A Canazart, Milton PS Ferreira, Carl R Andersson, Peter W Macfarlane, Wagner Meira Jr, et al. Automatic diagnosis of the 12-lead ecg using a deep neural network. Nature commu...

    2020

  75. [84]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High-resolution image synthesis with latent diffusion models. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022

    2022

  76. [85]

    A review of deep learning techniques for forecasting energy use in buildings

    Jason Runge and Radu Zmeureanu. A review of deep learning techniques for forecasting energy use in buildings. Energies, 14(3):608, 2021

    2021

  77. [86]

    Image- based time series forecasting: A deep convolutional neural network approach.Neural Networks, 157:39–53, 2023

    Artemios-Anargyros Semenoglou, Evangelos Spiliotis, and Vassilios Assimakopoulos. Image- based time series forecasting: A deep convolutional neural network approach.Neural Networks, 157:39–53, 2023

    2023

  78. [87]

    Dust: Dual swin transformer for multi- modal video and time-series modeling

    Liang Shi, Yixin Chen, Meimei Liu, and Feng Guo. Dust: Dual swin transformer for multi- modal video and time-series modeling. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4537–4546, 2024

    2024

  79. [88]

    Learning video representa- tions using contrastive bidirectional transformer

    Chen Sun, Fabien Baradel, Kevin Murphy, and Cordelia Schmid. Learning video representa- tions using contrastive bidirectional transformer. arXiv preprint arXiv:1906.05743, 2019

    1906 arXiv

  80. [89]

    Videobert: A joint model for video and language representation learning

    Chen Sun, Austin Myers, Carl V ondrick, Kevin Murphy, and Cordelia Schmid. Videobert: A joint model for video and language representation learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 7464–7473, 2019

    2019

  81. [90]

    Test: Text prototype aligned embedding to activate llm’s ability for time series

    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

  82. [91]

    Long-form video-language pre-training with multimodal temporal contrastive learning

    Yuchong Sun, Hongwei Xue, Ruihua Song, Bei Liu, Huan Yang, and Jianlong Fu. Long-form video-language pre-training with multimodal temporal contrastive learning. arXiv preprint arXiv:2210.06031, 2022

    2022 arXiv

  83. [92]

    Are language models actually useful for time series forecasting? In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    Mingtian Tan, Mike A Merrill, Vinayak Gupta, Tim Althoff, and Thomas Hartvigsen. Are language models actually useful for time series forecasting? In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  84. [93]

    Are language models actually useful for time series forecasting? In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    Mingtian Tan, Mike A Merrill, Vinayak Gupta, Tim Althoff, and Thomas Hartvigsen. Are language models actually useful for time series forecasting? In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. 15

    2024

  85. [94]

    Llama: Open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timo- thée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023

    2023 arXiv

  86. [95]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017

    2017

  87. [96]

    Towards data-centric ai: A com- prehensive survey of traditional, reinforcement, and generative approaches for tabular data transformation

    Dongjie Wang, Yanyong Huang, Wangyang Ying, Haoyue Bai, Nanxu Gong, Xinyuan Wang, Sixun Dong, Tao Zhe, Kunpeng Liu, Meng Xiao, et al. Towards data-centric ai: A com- prehensive survey of traditional, reinforcement, and generative approaches for tabular data transformation. arX...

    2025 arXiv

  88. [97]

    Micn: Multi-scale local and global context modeling for long-term series forecasting

    Huiqiang Wang, Jian Peng, Feihu Huang, Jince Wang, Junhui Chen, and Yifei Xiao. Micn: Multi-scale local and global context modeling for long-term series forecasting. In The eleventh international conference on learning representations, 2023

    2023

  89. [98]

    Long-short temporal contrastive learning of video transformers

    Jue Wang, Gedas Bertasius, Du Tran, and Lorenzo Torresani. Long-short temporal contrastive learning of video transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14010–14020, 2022

    2022

  90. [99]

    Actionclip: A new paradigm for video action recognition

    Mengmeng Wang, Jiazheng Xing, and Yong Liu. Actionclip: A new paradigm for video action recognition. arXiv preprint arXiv:2109.08472, 2021

    2021 arXiv

  91. [100]

    A hierarchal bert structure for native speaker writing detection

    Xinyuan Wang, Qinke Peng, Xu Mou, Haozhou Li, and Ying Wang. A hierarchal bert structure for native speaker writing detection. In 2022 China Automation Congress (CAC) , pages 3705–3710. IEEE, 2022

    2022

  92. [101]

    Lcmdc: Large-scale chinese medical dialogue corpora for automatic triage and medical consultation

    Xinyuan Wang, Haozhou Li, Dingfang Zheng, and Qinke Peng. Lcmdc: Large-scale chinese medical dialogue corpora for automatic triage and medical consultation. arXiv preprint arXiv:2410.03521, 2024

    2024 arXiv

  93. [102]

    Knockoff-guided feature selection via a single pre-trained reinforced agent

    Xinyuan Wang, Dongjie Wang, Wangyang Ying, Rui Xie, Haifeng Chen, and Yanjie Fu. Knockoff-guided feature selection via a single pre-trained reinforced agent. arXiv preprint arXiv:2403.04015, 2024

    2024 arXiv

  94. [103]

    Llm-enhanced user-item interactions: Leveraging edge information for optimized recommendations

    Xinyuan Wang, Liang Wu, Liangjie Hong, Hao Liu, and Yanjie Fu. Llm-enhanced user-item interactions: Leveraging edge information for optimized recommendations. arXiv preprint arXiv:2402.09617, 2024

    2024 arXiv

  95. [104]

    Llm-ml teaming: Integrated symbolic decoding and gradient search for valid and stable generative feature transformation

    Xinyuan Wang, Haoyue Bai, Nanxu Gong, Wangyang Ying, Sixun Dong, Xiquan Cui, and Yanjie Fu. Llm-ml teaming: Integrated symbolic decoding and gradient search for valid and stable generative feature transformation. arXiv preprint arXiv:2506.09085, 2025

    2025 arXiv

  96. [105]

    Mixllm: Dynamic routing in mixed large language models

    Xinyuan Wang, Yanchi Liu, Wei Cheng, Xujiang Zhao, Zhengzhang Chen, Wenchao Yu, Yanjie Fu, and Haifeng Chen. Mixllm: Dynamic routing in mixed large language models. arXiv preprint arXiv:2502.18482, 2025

    2025 arXiv

  97. [106]

    Efficient post-training refinement of latent reasoning in large language models

    Xinyuan Wang, Dongjie Wang, Wangyang Ying, Haoyue Bai, Nanxu Gong, Sixun Dong, Kunpeng Liu, and Yanjie Fu. Efficient post-training refinement of latent reasoning in large language models. arXiv preprint arXiv:2506.08552, 2025

    2025

  98. [107]

    Enhanced whole page optimization via mixed- grained reward mechanism-adapted language models

    Xinyuan Wang, Liang Wu, and Yanjie Fu. Enhanced whole page optimization via mixed- grained reward mechanism-adapted language models. arXiv preprint arXiv:2506.09084, 2025

    2025 arXiv

  99. [108]

    A successful hybrid deep learning model aiming at promoter identification

    Ying Wang, Qinke Peng, Xu Mou, Xinyuan Wang, Haozhou Li, Tian Han, Zhao Sun, and Xiao Wang. A successful hybrid deep learning model aiming at promoter identification. BMC bioinformatics, 23(Suppl 1):206, 2022

    2022

  100. [109]

    Deep time series models: A comprehensive survey and benchmark

    Yuxuan Wang, Haixu Wu, Jiaxiang Dong, Yong Liu, Mingsheng Long, and Jianmin Wang. Deep time series models: A comprehensive survey and benchmark. 2024

    2024

  101. [110]

    Timexer: Empowering transformers for time series forecasting with exogenous variables

    Yuxuan Wang, Haixu Wu, Jiaxiang Dong, Guo Qin, Haoran Zhang, Yong Liu, Yunzhong Qiu, Jianmin Wang, and Mingsheng Long. Timexer: Empowering transformers for time series forecasting with exogenous variables. arXiv preprint arXiv:2402.19072, 2024. 16

    2024 arXiv

  102. [111]

    Imaging time-series to improve classification and imputation

    Zhiguang Wang and Tim Oates. Imaging time-series to improve classification and imputation. arXiv preprint arXiv:1506.00327, 2015

    2015 arXiv

  103. [112]

    Electricity price forecasting: A review of the state-of-the-art with a look into the future

    Rafał Weron. Electricity price forecasting: A review of the state-of-the-art with a look into the future. International journal of forecasting, 30(4):1030–1081, 2014

    2014

  104. [113]

    Etsformer: Expo- nential smoothing transformers for time-series forecasting

    Gerald Woo, Chenghao Liu, Doyen Sahoo, Akshat Kumar, and Steven Hoi. Etsformer: Expo- nential smoothing transformers for time-series forecasting. arXiv preprint arXiv:2202.01381, 2022

    2022 arXiv

  105. [114]

    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, 34:22419–22430, 2021

    2021

  106. [116]

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

    Haixu Wu, Tengge Hu, Yong Liu, Hang Zhou, Jianmin Wang, and Mingsheng Long. Times- net: Temporal 2d-variation modeling for general time series analysis. arXiv preprint arXiv:2210.02186, 2022

    2022 arXiv

  107. [117]

    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(6):602–611, 2023

    2023

  108. [118]

    Wav2clip: Learning robust audio representations from clip

    Ho-Hsiang Wu, Prem Seetharaman, Kundan Kumar, and Juan Pablo Bello. Wav2clip: Learning robust audio representations from clip. In ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 4563–4567. IEEE, 2022

    2022

  109. [119]

    Deep multimodal representation learning from temporal data

    Xitong Yang, Palghat Ramesh, Radha Chitta, Sriganesh Madhvanath, Edgar A Bernal, and Jiebo Luo. Deep multimodal representation learning from temporal data. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5447–5455, 2017

    2017

  110. [120]

    Sichuan dialect speech recognition with deep lstm network

    Wangyang Ying, Lei Zhang, and Hongli Deng. Sichuan dialect speech recognition with deep lstm network. Frontiers of Computer Science, 14(2):378–387, 2020

    2020

  111. [121]

    Self-optimizing fea- ture generation via categorical hashing representation and hierarchical reinforcement crossing

    Wangyang Ying, Dongjie Wang, Kunpeng Liu, Leilei Sun, and Yanjie Fu. Self-optimizing fea- ture generation via categorical hashing representation and hierarchical reinforcement crossing. In 2023 IEEE International Conference on Data Mining (ICDM), pages 748–757. IEEE, 2023

    2023

  112. [122]

    Feature selection as deep sequential generative learning

    Wangyang Ying, Dongjie Wang, Haifeng Chen, and Yanjie Fu. Feature selection as deep sequential generative learning. ACM Transactions on Knowledge Discovery from Data, 18(9): 1–21, 2024

    2024

  113. [123]

    Unsupervised generative feature transformation via graph contrastive pre-training and multi-objective fine-tuning

    Wangyang Ying, Dongjie Wang, Xuanming Hu, Yuanchun Zhou, Charu C Aggarwal, and Yanjie Fu. Unsupervised generative feature transformation via graph contrastive pre-training and multi-objective fine-tuning. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery ...

    2024

  114. [124]

    Bridging the domain gap in equation distillation with reinforcement feedback

    Wangyang Ying, Haoyue Bai, Nanxu Gong, Xinyuan Wang, Sixun Dong, Haifeng Chen, and Yanjie Fu. Bridging the domain gap in equation distillation with reinforcement feedback. arXiv preprint arXiv:2505.15572, 2025

    2025 arXiv

  115. [125]

    A survey on data-centric ai: Tabular learning from reinforcement learning and generative ai perspective

    Wangyang Ying, Cong Wei, Nanxu Gong, Xinyuan Wang, Haoyue Bai, Arun Vignesh Malarkkan, Sixun Dong, Dongjie Wang, Denghui Zhang, and Yanjie Fu. A survey on data-centric ai: Tabular learning from reinforcement learning and generative ai perspective. arXiv preprint arXiv:2502.08828, 2025

    2025 arXiv

  116. [126]

    Large batch optimization for deep learning: Training bert in 76 minutes

    Yang You, Jing Li, Sashank Reddi, Jonathan Hseu, Sanjiv Kumar, Srinadh Bhojanapalli, Xiaodan Song, James Demmel, Kurt Keutzer, and Cho-Jui Hsieh. Large batch optimization for deep learning: Training bert in 76 minutes. arXiv preprint arXiv:1904.00962, 2019. 17

    1904 arXiv

  117. [127]

    Coca: Contrastive captioners are image-text foundation models

    Jiahui Yu, Zirui Wang, Vijay Vasudevan, Legg Yeung, Mojtaba Seyedhosseini, and Yonghui Wu. Coca: Contrastive captioners are image-text foundation models. arXiv preprint arXiv:2205.01917, 2022

    2022 arXiv

  118. [128]

    Are transformers effective for time series forecasting? In Proceedings of the AAAI conference on artificial intelligence , volume 37, pages 11121–11128, 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 , volume 37, pages 11121–11128, 2023

    2023

  119. [129]

    Sigmoid loss for language image pre-training

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. Sigmoid loss for language image pre-training. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11975–11986, 2023

    2023

  120. [130]

    Pointclip: Point cloud understanding by clip

    Renrui Zhang, Ziyu Guo, Wei Zhang, Kunchang Li, Xupeng Miao, Bin Cui, Yu Qiao, Peng Gao, and Hongsheng Li. Pointclip: Point cloud understanding by clip. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8552–8562, 2022

    2022

  121. [131]

    Less is more: Fast multivariate time series forecasting with light sampling-oriented mlp structures

    Tianping Zhang, Yizhuo Zhang, Wei Cao, Jiang Bian, Xiaohan Yi, Shun Zheng, and Jian Li. Less is more: Fast multivariate time series forecasting with light sampling-oriented mlp structures. arXiv preprint arXiv:2207.01186, 2022

    2022 arXiv

  122. [132]

    Integration of mamba and transformer-mat for long-short range time series forecasting with application to weather dynamics

    Wenqing Zhang, Junming Huang, Ruotong Wang, Changsong Wei, Wenqian Huang, and Yuxin Qiao. Integration of mamba and transformer-mat for long-short range time series forecasting with application to weather dynamics. In 2024 International Conference on Electrical, Communication a...

    2024

  123. [133]

    Skilful nowcasting of extreme precipitation with nowcastnet

    Yuchen Zhang, Mingsheng Long, Kaiyuan Chen, Lanxiang Xing, Ronghua Jin, Michael I Jordan, and Jianmin Wang. Skilful nowcasting of extreme precipitation with nowcastnet. Nature, 619(7970):526–532, 2023

    2023

  124. [134]

    Crossformer: Transformer utilizing cross-dimension depen- dency for multivariate time series forecasting

    Yunhao Zhang and Junchi Yan. Crossformer: Transformer utilizing cross-dimension depen- dency for multivariate time series forecasting. In The eleventh international conference on learning representations, 2023

    2023

  125. [135]

    Time-vlm: Exploring multimodal vision-language models for augmented time series forecasting, 2025

    Siru Zhong, Weilin Ruan, Ming Jin, Huan Li, Qingsong Wen, and Yuxuan Liang. Time-vlm: Exploring multimodal vision-language models for augmented time series forecasting, 2025. URL https://arxiv.org/abs/2502.04395

    2025 arXiv

  126. [136]

    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, volume 35, pages 11106–11115, 2021

    2021

  127. [138]

    Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting

    Tian Zhou, Ziqing Ma, Qingsong Wen, Xue Wang, Liang Sun, and Rong Jin. Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting. InInternational conference on machine learning, pages 27268–27286. PMLR, 2022

    2022

  128. [140]

    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 , 36:43322– 43355, 2023

    2023

  129. [141]

    Actbert: Learning global-local video-text representations

    Linchao Zhu and Yi Yang. Actbert: Learning global-local video-text representations. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8746–8755, 2020. 18 A Implementation Details For model initialization, we adopt Kaiming uniform initi...

    2020

  130. [2025]

    URL https://doi.org/10.20944/ preprints202501.1003.v3

    doi: 10.20944/preprints202501.1003.v3. URL https://doi.org/10.20944/ preprints202501.1003.v3

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.