UniTime: A Language-Empowered Unified Model for Cross-Domain Time Series Forecasting
Open accessThe DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Abstract
Multivariate time series forecasting plays a pivotal role in contemporary web technologies. In contrast to conventional methods that involve creating dedicated models for specific time series application domains, this research advocates for a unified model paradigm that transcends domain boundaries. However, learning an effective cross-domain model presents the following challenges. First, various domains exhibit disparities in data characteristics, e.g., the number of variables, posing hurdles for existing models that impose inflexible constraints on these factors. Second, the model may encounter difficulties in distinguishing data from various domains, leading to suboptimal performance in our assessments. Third, the diverse convergence rates of time series domains can also result in compromised empirical performance. To address these issues, we propose UniTime for effective cross-domain time series learning. Concretely, UniTime can flexibly adapt to data with varying characteristics. It also uses domain instructions and a Language-TS Transformer to offer identification information and align two modalities. In addition, UniTime employs masking to alleviate domain convergence speed imbalance issues. Our extensive experiments demonstrate the effectiveness of UniTime in advancing state-of-the-art forecasting performance and zero-shot transferability.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Foundation Models learn general-purpose representations from large-scale data that can be adapted to diverse downstream tasks [6,8,13,19,25,28,30,35,45,60,62,69].”
From Towards Customizable Foundation Models for Human Activity Recognition with Wearable Devices (2025) · cites this paper as [25]
Citations
Cites
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., & Guo, B. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00986 · Cited by 35,534 · cited 1×
“Recently, Transformers [28] have achieved exceptional performance in various tasks of natural language processing [14, 25, 26] and computer vision [3, 5, 21], which also triggered significant interest in the time series community [31].”
Background(2008) · Journal of Machine Learning Research · Cited by 33,051
End-to-End Object Detection with Transformers
Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58452-8_13 · Cited by 18,285
End-to-End Object Detection with Transformers
Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020) · arXiv · doi:10.48550/arxiv.2005.12872 · Cited by 18,253 · cited 1×
“XXXXXXX and computer vision [3, 5, 21], which also triggered significant interest in the time series community [31].”
BackgroundKirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W.-Y., et al. (2023) · 2023 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00371 · Cited by 11,953 · cited 1×
“Such a unified model paradigm has achieved remarkable success in computer vision [15, 23], natural language processing [1, 25], and holds promise in the context of time series modeling.”
BackgroundInformer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting
Zhou, H., Zhang, S., Peng, J., Zhang, S., Li, J., Xiong, H., & Zhang, W. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i12.17325 · Cited by 7,752
Are Transformers Effective for Time Series Forecasting?
Zeng, A., Chen, M., Zhang, L., & Xu, Q. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i9.26317 · Cited by 3,320
Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks
Lai, G., Chang, W.-C., Yang, Y., & Liu, H. (2018) · The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval · doi:10.1145/3209978.3210006 · Cited by 2,152 · cited 2×
“For example, ETTm2, ETTh2, Exchange, and Illness experience severe overfitting beyond the 4th epoch, while others require more epochs to reach convergence.”
BackgroundSelf-Instruct: Aligning Language Models with Self-Generated Instructions
Wang, Y., Kordi, Y., Mishra, S., Liu, A., Smith, N. A., Khashabi, D., & Hajishirzi, H. (2023) · Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2023.acl-long.754 · Cited by 1,243
An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks
Goodfellow, I. J., Mirza, M., Xiao, D., Courville, A., & Bengio, Y. (2013) · arXiv · doi:10.48550/arxiv.1312.6211 · Cited by 1,089 · cited 1×
“However, this method often results in unstable learning and the issue of catastrophic forgetting [6].”
Method
Research works printed but not linked: 24
- Tom Brown Benjamin Mann Nick Ryder Melanie Subbiah Jared D Kaplan Prafulla Dhariwal Arvind Neelakantan Pranav Shyam Girish Sastry Amanda Askell et al. 2020. Language models are few-shot learners. In Advances in neural information processing systems. 1877--1901.
- Shizhe Diao, Rui Pan, Hanze Dong, Ka Shun Shum, Jipeng Zhang, Wei Xiong, and Tong Zhang. 2023. Lmflow: An extensible toolkit for finetuning and inference of large foundation models. arXiv preprint arXiv:2306.12420 (2023). · arXiv
- Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. 2021. An image is worth 16x16 words: Transformers for image recognition at scale. In International Conference on Learning Representations.
- Nate Gruver Marc Finzi Shikai Qiu and Andrew Gordon Wilson. 2023. Large language models are zero-shot time series forecasters. In Advances in Neural Information Processing Systems.
- Xiaoxin He, Xavier Bresson, Thomas Laurent, and Bryan Hooi. 2023. Explanations as Features: LLM-Based Features for Text-Attributed Graphs. arXiv preprint arXiv:2305.19523 (2023). · arXiv
- Yupeng Hou, Junjie Zhang, Zihan Lin, Hongyu Lu, Ruobing Xie, Julian McAuley, and Wayne Xin Zhao. 2023. Large language models are zero-shot rankers for recommender systems. arXiv preprint arXiv:2305.08845 (2023). · arXiv
- Ming Jin Qingsong Wen Yuxuan Liang Chaoli Zhang Siqiao Xue Xue Wang James Zhang Yi Wang Haifeng Chen Xiaoli Li et al. 2023. Large models for time series and spatio-temporal data: A survey and outlook. arXiv preprint arXiv:2310.10196 (2023). · arXiv
- Ming Jin, Yifan Zhang, Wei Chen, Kexin Zhang, Yuxuan Liang, Bin Yang, Jindong Wang, Shirui Pan, and Qingsong Wen. 2024. Position Paper: What Can Large Language Models Tell Us about Time Series Analysis. arXiv preprint arXiv:2402.02713 (2024). · arXiv
- Jacob Devlin Ming-Wei Chang Kenton and Lee Kristina Toutanova. 2019. Bert: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of NAACL-HLT. 2.
- Nikita Kitaev, Łukasz Kaiser, and Anselm Levskaya. 2020. Reformer: The efficient transformer. In International Conference on Learning Representations.
- Shiyang Li Xiaoyong Jin Yao Xuan Xiyou Zhou Wenhu Chen Yu-Xiang Wang and Xifeng Yan. 2019. Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. In Advances in neural information processing systems.
- Shizhan Liu, Hang Yu, Cong Liao, Jianguo Li, Weiyao Lin, Alex X Liu, and Schahram Dustdar. 2022b. Pyraformer: Low-complexity pyramidal attention for long-range time series modeling and forecasting. In International Conference on Learning Representations.
- Yong Liu, Haixu Wu, Jianmin Wang, and Mingsheng Long. 2022a. Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting. In Advances in Neural Information Processing Systems. 9881--9893.
- Jun Ma and Bo Wang. 2023. Segment anything in medical images. arXiv preprint arXiv:2304.12306 (2023). · arXiv
- Yuqi Nie, Nam H Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. 2023. A time series is worth 64 words: Long-term forecasting with transformers. In International Conference on Learning Representations.
- Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. The Journal of Machine Learning Research (2020), 5485--5551.
- Ashish Vaswani Noam Shazeer Niki Parmar Jakob Uszkoreit Llion Jones Aidan N Gomez Łukasz Kaiser and Illia Polosukhin. 2017. Attention is all you need. In Advances in neural information processing systems. 5998--6008.
- Gerald Woo, Chenghao Liu, Doyen Sahoo, Akshat Kumar, and Steven Hoi. 2022. Etsformer: Exponential smoothing transformers for time-series forecasting. In arXiv preprint arXiv:2202.01381. · arXiv
- Haixu Wu, Tengge Hu, Yong Liu, Hang Zhou, Jianmin Wang, and Mingsheng Long. 2023. Timesnet: Temporal 2d-variation modeling for general time series analysis. In International Conference on Learning Representations.
- Junjie Zhang, Ruobing Xie, Yupeng Hou, Wayne Xin Zhao, Leyu Lin, and Ji-Rong Wen. 2023. Recommendation as instruction following: A large language model empowered recommendation approach. arXiv preprint arXiv:2305.07001 (2023). · arXiv
- Yunhao Zhang and Junchi Yan. 2022. Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting. In The Eleventh International Conference on Learning Representations.
- Haiteng Zhao, Shengchao Liu, Chang Ma, Hannan Xu, Jie Fu, Zhi-Hong Deng, Lingpeng Kong, and Qi Liu. 2023. GIMLET: A Unified Graph-Text Model for Instruction-Based Molecule Zero-Shot Learning. In Advances in neural information processing systems.
- Tian Zhou, Ziqing Ma, Qingsong Wen, Xue Wang, Liang Sun, and Rong Jin. 2022. Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting. In International Conference on Machine Learning. 27268--27286.
- Tian Zhou Peisong Niu Xue Wang Liang Sun and Rong Jin. 2023. One Fits All: Power General Time Series Analysis by Pretrained LM. In Advances in Neural Information Processing Systems.
Cited by
Foundation Models for Time Series Analysis: A Tutorial and Survey
Liang, Y., Wen, H., Nie, Y., Jiang, Y., Jin, M., Song, D., Pan, S., & Wen, Q. (2024) · Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining · doi:10.1145/3637528.3671451 · Cited by 225
Self-citationDeep learning for cross-domain data fusion in urban computing: Taxonomy, advances, and outlook
Zou, X., Yan, Y., Hao, X., Hu, Y., Wen, H., Liu, E., Zhang, J., Li, Y., Li, T., Zheng, Y., et al. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102606 · Cited by 176
Self-citationEmpowering Time Series Analysis with Large Language Models: A Survey
Jiang, Y., Pan, Z., Zhang, X., Garg, S., Schneider, A., Nevmyvaka, Y., & Song, D. (2024) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2024/895 · Cited by 145
ST-LLM+: Graph Enhanced Spatio-Temporal Large Language Models for Traffic Prediction
Liu, C., Hettige, K. H., Xu, Q., Long, C., Xiang, S., Cong, G., Li, Z., & Zhao, R. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3570705 · Cited by 116
“Parameter-tuning approaches modify LLMs to suit time series tasks [30], [31], while other methods preprocess time series data into formats like textual prompts or use reprogramming to fit the LLMs input space [32], [33].”
CMLLM: A novel cross-modal large language model for wind power forecasting
Zhu, G., Jia, W., Xing, Z., Xiang, L., Hu, A., & Hao, R. (2025) · Energy Conversion and Management · doi:10.1016/j.enconman.2025.119673 · Cited by 64
Multi-modal Time Series Analysis: A Tutorial and Survey
Jiang, Y., Ning, K., Pan, Z., Shen, X., Ni, J., Yu, W., Schneider, A., Chen, H., Nevmyvaka, Y., & Song, D. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3711896.3736567 · Cited by 83
“To effectively leverage cross-modal information, existing methods often incorporate alignment designs after concatenating representations [5, 11, 12, 18, 31, 33, 41, 42, 44, 54, 55, 60, 69, 70, 74, 85, 96].”
InfluentialPeriodformer: An efficient long-term time series forecasting method based on periodic attention
Liang, D., Zhang, H., Yuan, D., & Zhang, M. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112556 · Cited by 36
Cheng, H., Gong, Z., & Wang, C. (2025) · Applied Soft Computing · doi:10.1016/j.asoc.2025.113174 · Cited by 41
TimeRAF: Retrieval-Augmented Foundation Model for Zero-Shot Time Series Forecasting
Zhang, H., Xu, C., Zhang, Y., Zhang, Z., Wang, L., & Bian, J. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3579137 · Cited by 67
“UniTime[22]introducesauni-fiedmodelthatleveragesdomaininstructionsandalanguage– timeseriestransformertosupportzero-shotandcross-domain timeseriesforecasting.”
LLMs and IoT: A Comprehensive Survey on Large Language Models and the Internet of Things
Sarhaddi, F., Nguyen, N. T., Zuniga, A., Khan, M., Hui, P., Tarkoma, S., Flores, H., & Nurmi, P. (2026) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2026.3680064 · Cited by 33
“These developments have enabled LLMs to better handle various IoT sensor data types, such as spatio-temporal data [241], [242], raw IMU data [99], [101], [243], sensor data from smart homes [105], and time-series data [244], [245], [246], [247], [248], [249].”
Deep anomaly detection for time series: A survey
Jia, X., Peng, X., Peng, W., Zhao, B., Li, H., & Shen, C. (2025) · Computer Science Review · doi:10.1016/j.cosrev.2025.100787 · Cited by 37
Foundation Models for Spatio-Temporal Data Science: A Tutorial and Survey
Liang, Y., Wen, H., Xia, Y., Jin, M., Yang, B., Salim, F. D., Wen, Q., Pan, S., & Cong, G. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3711896.3736552 · Cited by 47
Self-citationLiu, C., Hao, M., Xu, Q., Zhou, S., Long, C., Zhao, Y., Li, Z., & Zhao, R. (2025) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde65448.2025.00237 · Cited by 41
“UniTime introduces a unified model that leverages natural language prompts and a Language-TS Transformer to align domain-specific characteristics for effective cross-domain time series forecasting [49].”
InfluentialZhang, T., Yu, H., Wang, X., Wang, J., & Ren, B. (2025) · Automation in Construction · doi:10.1016/j.autcon.2025.106036 · Cited by 18
Prompt-Based Spatio-Temporal Graph Transfer Learning
Hu, J., Liu, X., Fan, Z., Yin, Y., Xiang, S., Ramasamy, S., & Zimmermann, R. (2024) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3627673.3679554 · Cited by 37
Self-citationAre Time Series Foundation Models Ready to Revolutionize Predictive Building Analytics?
Mulayim, O. B., Quan, P., Han, L., Ouyang, X., Hong, D., Berges, M. E., & Srivastava, M. B. (2024) · ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (BuildSys) · doi:10.1145/3671127.3698177 · Cited by 20
“Models like UniTime [12] and TimeLLM [11] incorporate context but are not zero-shot predictors.”
Zhang, L., Cheng, W., Zhang, S., Xing, J., Nie, Z., Chen, X., Lan, D., Liu, Y., Yang, Y., & Pang, Z. (2025) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2025.3575118 · Cited by 19
“UniTime [61] leverages domain knowledge as prompts for greater flexibility, while GAT-GPT [62] and spatial-temporal large language model (ST-LLM) [63] incorporate graph attention and spatiotemporal embeddings to handle complex dependencies.”
A channel-independent network based on wavelet enhancement for long-term time series forecasting
Huang, Z., Zhang, F., & Liu, Y. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.110964 · Cited by 12
GLALLM: Adapting LLMs for spatio-temporal wind speed forecasting via global–local aware modeling
Wu, T., & Ling, Q. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113739 · Cited by 14
Li, J., Li, D., Yang, Y., Xi, H., Xiao, Y., Xiao, Y., & Yang, Q. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.131288 · Cited by 12
A Survey of Large Language Models for Traffic Forecasting: Methods and Applications
Long, Q., Liu, S., Cao, N., Ren, Z., Luo, X., Ju, W., Fang, C., Zhu, Z., Zhu, H., & Zhou, Y. (2026) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2026.3668685 · Cited by 25
“We select five representative works [100], [141], [142], [143], [144], [145] to conduct experiments on a popularly-used dataset [146].”
InstrucTime: Advancing Time Series Classification with Multimodal Language Modeling
Cheng, M., Chen, Y., Liu, Q., Liu, Z., Luo, Y., & Chen, E. (2025) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3701551.3703499 · Cited by 39
Time-Series Forecasting With Semantic-Aware Trend Loss Function
Liao, H., Liu, X., Guo, L., Li, Y., & Chen, Q. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3651732 · Cited by 9
“The first ones are the NLP conversion of time series [15], [18], which mainly involves converting time-series data into a format suitable for processing by NLP models.”
Self-citationExploiting Language Power for Time Series Forecasting with Exogenous Variables
Huang, Q., Zhou, Z., Yang, K., & Wang, Y. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714793 · Cited by 28
“It is difficult to model such changing and diverse impact only by observed numeric [20], highlighting the necessity of thoroughly learning multi-grained temporal features to effectively model these intricate patterns [14].”
InfluentialHan, S., Liang, Z., & Li, S. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.115155 · Cited by 10
Can time-series foundation models perform building energy management tasks?
Mulayim, O. B., Quan, P., Han, L., Ouyang, X., Hong, D., Berges, M. E., & Srivastava, M. B. (2026) · Data-Centric Engineering · doi:10.1017/dce.2026.10040 · Cited by 9
“Uni-Time (Liu et al., 2024) uses an encoder-based transformer, incorporating semantic instructions through a language encoder to handle domain confusion.”
Gong, J., Ding, J., Meng, F., Chen, G., Chen, H., Zhao, S., Lu, H., & Li, Y. (2024) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3637528.3671984 · Cited by 7
“In this regard, PLMs have been successfully adapted to other cross-domain tasks related to human behaviors, like recommender systems [2, 7, 45] and mobility trajectories [11, 20, 38].”
Self-citationHierarchical Multimodal LLMs with Semantic Space Alignment for Enhanced Time Series Classification
Tao, X., Pan, T., Cheng, M., Luo, Y., Liu, Q., & Chen, E. (2025) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3785505 · Cited by 14
Zhang, Z., Wang, W., Zhou, X., Bai, Y., & Yin, H. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3619521 · Cited by 6
“Additionally, pre-trained models based on large language models [9], [52], [53] encode a wide range of data features from diverse internet patterns into the model’s hidden layers, guiding downstream prediction tasks.”
Liu, Y., Huang, Z., Zhang, F., & Zhang, X. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.126584 · Cited by 6
Wang, P., Zheng, H., Xu, Q., Dai, S., Wang, Y., Yue, W., Zhu, W., Qian, T., & Zhao, L. (2025) · Symmetry · doi:10.3390/sym17030401 · Cited by 15
Metadata-informed transformers for informative time series forecasting
Dong, J., Wu, H., Wang, Y., Zhang, L., Wang, J., & Long, M. (2025) · Scientia Sinica Informationis · doi:10.1360/ssi-2024-0397 · Cited by 13
BALM-TSF: Balanced Multimodal Alignment for LLM-Based Time Series Forecasting
Zhou, S., Schöner, H., Lyu, H., Fouché, E., & Wang, S. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3761278 · Cited by 15
A Survey of Deep Learning for Time Series Forecasting: Taxonomy, Analysis and Future Directions
Guo, Q., Zhao, B., Song, M., & Zhong, G. (2026) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2026.3671378 · Cited by 9
“UniTime [112] leverages the power of language models to construct a universal framework capable of adapting to time series data across diverse domains.”
Towards Customizable Foundation Models for Human Activity Recognition with Wearable Devices
Qiu, M., Weng, C., Fan, M., & Wu, K. (2025) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3749479 · Cited by 4
“Foundation model have been pivotal in predicting signals on the different downstream classification tasks,across domains such as energy and finance, emphasizing common signal properties [6, 8, 13, 19, 25, 30, 35, 60, 69].”
Revisiting Backdoor Attacks on Time Series Classification in the Frequency Domain
Huang, Y., Zhang, M., Wang, Z., Li, W., & Yang, M. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714827 · Cited by 12
Sun, Y., Wang, X., Cao, G., & Mao, S. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3604332 · Cited by 5
“For example, UmiTime [14] shows that LLMs can significantly enhance time series prediction, particularly in contexts involving limited data or cross-domain challenges.”
Visible light human activity recognition driven by generative language model
Yang, Y., Liu, Z., Chen, Y., Yan, B., Sun, Y., & Feng, T. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103159 · Cited by 4
Digital Twin for Capacity Estimation of Lithium-Ion Batteries With Large Language Models
Qin, D., & Yang, J. H. (2025) · IEEE Transactions on Transportation Electrification · doi:10.1109/tte.2025.3645866 · Cited by 4
He, P., Gan, Y., Liu, Y., Lin, R., Zhou, G., Xie, J., Liu, Y., & Liu, Q. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.115557 · Cited by 5
Advancing Financial Engineering with Foundation Models: Progress, Applications, and Challenges
Chen, L., Liu, S., Yan, J., Wang, X., Liu, H., Li, C., Jiao, K., Ying, J., Liu, Y., Yang, Q., et al. (2025) · Engineering · doi:10.1016/j.eng.2025.11.029 · Cited by 15
“In contrast, UniTime [86], a fine-tuned GPT-2 variant, updates the LLM parameters directly using multivariate time-series data and domain-specific instruction tuning, thereby enabling it to model temporal dependencies natively.”
CLaSP: Learning Concepts for Time-Series Signals from Natural Language Supervision
Ito, A., Dohi, K., & Kawaguchi, Y. (2025) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco63237.2025.11226094 · Cited by 6
EVCharging-GPT: Predicting Electric Vehicle User Charging Behavior Using Large Language Models
Li, H., Wang, J., Jiang, Z., Su, G., Yan, C., Su, H., & Wang, Z. (2025) · Data Intelligence · doi:10.3724/2096-7004.di.2025.0002 · Cited by 4
PulseFormer: Continuous Remote Heart Rate Measurement Through Zoomed Time-Spectral Attention
Comas, J., Ruiz, A., & Sukno, F. M. (2025) · IEEE Transactions on Biometrics Behavior and Identity Science · doi:10.1109/tbiom.2025.3544647 · Cited by 2
“The introduction of the Transformer architecture [72] has significantly impacted not only natural language processing (NLP) [5], [29] but also a wide range of other fields such as robotics [4], [64] and time series forecasting [26], [27], [39].”
Context-driven cold-start Web traffic forecasting
Zhou, X., Wang, W., Buntine, W., & Bergmeir, C. (2025) · World Wide Web · doi:10.1007/s11280-025-01359-7 · Cited by 2
DeMa: dual-path delay-aware Mamba for efficient multivariate time series analysis
An, R., Qu, H., Fan, W., Shang, X., & Li, Q. (2026) · Frontiers of Computer Science · doi:10.1007/s11704-026-52221-6 · Cited by 1
“Representative approaches include Informer [20], Fedformer [21], Auto-former [22], Pyraformer [23], PatchTST [8], Crossformer [24], and UniTime [63].”
Preference Guided Meta-Learning for Cross Domain Time Series Forecasting
Li, X., Teng, F., Li, T., & Duan, Q. (2026) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2026.3658637 · Cited by 3
“To ensure a fair comparison, we adhere to the experimental settings established by Liu et al.[3].”
InfluentialLLM-ABBA: Understanding Time Series via Symbolic Approximation
Chen, X., Carson, E., & Cheng, K. (2026) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2026.3662011 · Cited by 3
“As studied in [18], the ABBA forecasting framework offers several advantages over directly applying LSTMs to raw time series data: (i) the method effectively compresses the series, removes noise, and captures the essential structural patterns.”
Enhancing Large Language Models for Time-Series Forecasting via Vector-Injected In-Context Learning
Zhang, J., Wang, J., Qiang, W., Xu, F., & Zheng, C. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774904.3792374 · Cited by 6
Tabassum, S., & Nower, N. (2026) · PLoS ONE · doi:10.1371/journal.pone.0349527 · Cited by 2
LLM4cast: Repurposed LLM for Viral Disease Forecasting
Saeed, F., Aldosari, M., Arpinar, I. B., & Miller, J. A. (2024) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata62323.2024.10826009 · Cited by 6
Xi, X., Zha, R., He, C., Li, M., Yu, L., & Zhao, W. (2025) · Journal of Management Analytics · doi:10.1080/23270012.2025.2505442 · Cited by 2
Uncertainty-aware multi-modal time series anomaly detection via reduced-order evidence fusion
Li, Q., Wan, X., Guo, H., (17594), W. W., (261774), M. L., & Cui, Y. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.114212 · Cited by 1
Long-term time series forecasting based on Siamese network: a perspective on few-shot learning
Fan, J., Xiang, J., Liu, J., Wang, Z., & Wu, H. (2024) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-024-02317-x · Cited by 2
Web-Centric Human Mobility Analytics: Methods, Applications, and Future Directions in the LLM Era
Zhang, Z., Miao, H., Liang, Y., Zhao, Y., Han, X., Jia, P., Yang, B., & Jensen, C. S. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3701716.3715855 · Cited by 1
Self-citationKnowledge-enhanced AI models for domain-specific application: a survey
Li, X.-Y., Wang, Y., Liang, J., Wang, X., Zhou, P., Zhao, Q., Zhang, Y., & Song, Q. (2026) · Science China Information Sciences · doi:10.1007/s11432-025-4635-0 · Cited by 1
FedRMamba: Federated Residual Mamba for Multivariate Time-Series Forecasting
Hu, Z., Zhang, L., & Zhu, G. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774904.3792712 · Cited by 1
Multimodal Large Language Model-Enabled Robust Beamforming for HAP Downlink Communications
Xing, X., Yang, P., Tao, G., Lu, D., Xiong, Z., & Cao, X. (2026) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2026.3693275 · Cited by 1
“In the time-series domain, recent works [32], [33], [34], [35] show that LLMs can be repurposed for forecasting when inputs are appropriately encoded.”
Yue, J., Zhao, C., & Cheng, X. (2026) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2026.3706386 · Cited by 1
Zhang, L.-Z., Zhai, Y.-P., Jia, T., He, M.-H., Duan, C.-M., Liu, Z.-Y., Ding, B.-L., & Li, Y. (2026) · SIGSOFT FSE Companion · doi:10.1145/3803437.3805206 · Cited by 8
Time Series Analysis in the Web: Recent Advances and Future Trends
Jin, M., Jin, W., Xue, H., Salim, F. D., & Liang, Y. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3701716.3715853 · Cited by 1
Self-citationAbductive learning-guided uncertainty modeling for time series anomaly detection
Zhang, Q., Zhu, M., Li, J., Bao, J., Zhang, D., & Chen, L. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113878 · Cited by 1
A Wind Power Forecasting Method Based on Large Language ModelsPreprint
Liu, D., & 橘高, 義. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5251506 · Cited by 1
LLM-ABBA: Understanding Time Series via Symbolic ApproximationPreprint
Carson, E., Chen, X., & Kang, C. (2025) · Qeios · doi:10.32388/wd5bow
Exploring Inter-Variate and Long-Term Dependencies to Boost Multivariate Time Series Forecasting
Ding, X. L., He, Y., Zhou, S., Liu, G., & Zhou, Q. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890320
“Their performance results in Table I are from [14].”
InfluentialExplorations in How Large Language Models Understand Time SeriesPreprint
Niu, H., Xiong, Y., Bao, J., Yuan, C., Yang, B., Xie, Y., Qiu, T., Li, M., & Tang, C. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5263703
TDSR: Temporal Dynamics Enhanced Semantic Recommendation via Cross-Domain Interactive Learning
Zhu, X., Song, Y., Guo, X., Wang, J., & Du, J. (2025) · Communications in computer and information science · doi:10.1007/978-981-95-0017-8_11
Foundation Model for Degradation Process Prediction of Rotating MachineryPreprint
Liu, X., Li, T., Chen, Z., Zhang, Z., Zhang, J., & Si, X. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5397401
Self-citationMABO: A Parallel Hyperparameter Search Algorithm for Time Series ForecastingPreprint
Liang, D., Chen, J., Zhang, H., & Wang, Y. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5458383
Chu, J., Liu, C., Bai, X., & Tan, J. (2025) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2025.3631696
“Some LLM-based MTS forecasting models [8], [11], [12] summarize textual knowledge prompts, such as domain background and dataset statistics, and directly concatenate them with MTS embeddings to inform domain-specific prediction.”
InfluentialZhou, Y., Xia, K., Liu, T., Wang, S., & Deng, X. (2025) · IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA) · doi:10.1109/ispa67752.2025.00093
“2) Baselines and Metrics: We include eight multivariate time series forecasting methods for comparison, which are categorized as follows: (1) LLM-based time series forecasting methods: TimeCMA [18], Time-LLM [26], UniTime [16], and OFA [17].”
Fiscal and Tax QA via Clause-Level Knowledge Graphs and Retrieval-augmented Large Language Models
Chen, H., Bu, X., & Tian, S. (2025) · International Conference on Computer Information Science and Artificial Intelligence (CISAI) · doi:10.1145/3773365.3773416
MABO: A Parallel Hyperparameter Search Algorithm for Time Series ForecastingPreprint
Liang, D., Chen, J., zhang, H., & Wang, Y. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6110247
Causal Graph Fuzzy LLMs: A First Introduction and Applications in Time Series Forecasting
Orang, O., Lucas, P. D. O., de Paiva, G. I. F., Silva, P. C. D. L. E., Silva, F. A. R., Veloso, A. A., & Guimarães, F. B. A. P. (2025) · Congresso Brasileiro de Inteligência Computacional (CBIC) · doi:10.21528/cbic2025-1187477 · Cited by 2
Cross-Modal Adaptation of Large Language Models for Building Energy Forecasting
Gu, Y., & Jazizadeh, F. (2026) · ASCE International Conference on Computing in Civil Engineering (i3CE) · doi:10.1061/9780784486436.061
ForeT2F: Enhancing Zero-Shot Prediction via Two-Stage Temporal-Frequency PretrainingPreprint
Shen, P., Xu, L., Xu, P., Liu, Y., & Li, H. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6170129
HTHTA- ViT ++: Edge-Optimised Hierarchical Transformer for ECG and Smart-Grid Time-Series
Khurmi, S., Yadav, N., Sharma, P., Arora, V., Tomar, A., Sharma, K., & Kumar, B. N. (2025) · IEEE Pune Section International Conference (PuneCon) · doi:10.1109/punecon67554.2025.11378571
He, Q., Zhu, X., Zhou, Y., Ding, F., & Chen, W. (2026) · International Conference on Intelligent Control and Information Processing · doi:10.1109/icicip67436.2026.11417616
“To further solve the problem of scarce financial data, Liu et al. constructed UniTime [30].”
Chen, Y., Zhao, R., Xu, R., Peng, J., & Qi, F. (2025) · IEEE Conference on Energy Internet and Energy System Integration (EI2) · doi:10.1109/ei268505.2025.11425347
“The second approach retrains partial encoder-decoder parameters on specific datasets [4, 5, 6], achieving good few-shot learning capabilities.”
Wang, H., Yu, S., & Liu, W. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6538785
DEEP TPC: Temporal-Prior Conditioning for Time Series Forecasting
Bellos, F., Premkumar, N., Avrithis, Y., Nguyen, N. H., & Corso, J. J. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11460749
Hou, R. L., Liu, Y., Wang, J., Liu, Q., Zhang, L., & Zhou, W. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463046
Patch-Aware Decomposition and Dynamic Fusion Network for Multivariate Time Series Forecasting
Zhou, J., Yu, W., Guo, W., & Zhang, J. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463460
TPP-LLM: Time Series Popularity Prediction via LLM Empowered by Textual Prototype and Prompt
Wang, Z., Xie, Z., Sun, Y. W., Niu, W., & Su, C. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464229
LLM-Augmented Stiefel Graph Neural Networks for Zero-Shot Spatio-Temporal Forecasting
郑建凯, Xie, L., Zhu, L., & Yang, G. (2026) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2026.3689935
Cart: A Future-Oriented Text Impacts Alignment Framework for Time Series Forecasting
Gu, H., Xiong, H., Huang, K., Zhu, W., Dai, Y., He, Z., Jing, Y., Zhang, K., & Wang, X. S. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-0372-7_15
A Controllable Agentic Framework for Time-Series Data Synthesis in Few-Shot Scenarios
Chen, T., Wang, S., Wu, H., & Xiong, R. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-0372-7_33
Mao, D., Li, S., Zhang, G., Wang, W., & Wang, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-0372-7_32
LogicGate: Adaptive Rule-Based Modeling of Exogenous Effects for Time Series Forecasting
Zhong, J., Chen, T., Yuan, W., Cui, L., & Yin, H. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-0375-8_15 · Cited by 1
Learning a Causation-Driven Retrieval Model for Effective Time Series Augmentation and Forecasting
Zhang, B., & Luo, D. (2026) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2026.3694186
“Several methods integrate textual descriptions as prompts to improve time series forecasting performance [8], [9], [10].”
LLM-Enhanced Web-Centric Spatio-Temporal Intelligence: Methods, Applications, and Frontier Research
Zhang, Z., Miao, H., Liang, Y., Zhao, Y., & King, I. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774905.3793915
Self-citationZhang, Q., Xu, Z., Cen, C., Zou, Y., Liu, H., Deng, Y., Wu, H., Liao, Z., Lin, Z., Li, D., et al. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774905.3795461
Zhang, W., Zheng, L., Zhang, J., Shang, J., Liu, W. T., & Zhong, J. (2026) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2026.3692535
“This has spurred growing interest in leveraging pretrained LLMs for time series analysis [15], [18], [19], [20], [21], [22], [23].”
Schema-Agnostic Feature Extractor Using LLM and Transformer for RUL Prediction
Takayama, R., & Natsumeda, M. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-1462-4_25
Toward In-Depth Root Cause Localization for Microservices With Multi-Agent Recursion-of-Thought
Zhang, L., Jia, T., Wang, K., Duan, C., He, M., Wang, R., Peng, X., Wang, M., Zhang, G., Chen, R., et al. (2026) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2026.3706572 · Cited by 4
“Similarly, ShellGPT [78] trains a model capable of automatically generating shell scripts for automated mitigation.”
Self-citationTimeNLP: Efficient Time Series Captioning via Decoupled Latent Semantic Alignment
Zhu, K., Wang, Y. M., & Zhang, L. (2026) · International Conference on Smart Grid and Artificial Intelligence (SGAI) · doi:10.1109/sgai69026.2026.11565826
“The remarkable reasoning and generalization capabilities of Large Language Models (LLMs) have sparked growing interest in adapting them for time series analysis [8], [10], [13].”
Li, X. F., Zheng, Y., & Wei, Z. (2026) · Frontiers of Computer Science · doi:10.1007/s11704-026-50462-z
Shao, Z., Yang, R., Dou, Y., Ye, B., Wu, B., & Zhang, G. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3384-7_41
Zhang, Z., Xiao, C., Ma, J., Zhang, T., Gao, Y., & Qian, H. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3485-1_25
Jie, W., Ma, D., Junkai, W., Wang, Q. Y., & Zhang, J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7190903
Metadata
- Authors
- Liu, X., Hu, J., Li, Y., Diao, S., Liang, Y., Hooi, B., & Zimmermann, R.
- Affiliations
- National University of Singapore (SG)
- Hong Kong University of Science and Technology (HK)
- The Hong Kong University of Science and Technology (Guangzhou) (CN)
- Venue
- Proceedings of the ACM Web Conference 2024
- Publisher
- ACM
- Published
- 2024-05-13
- Pages
- 4095-4106
- OpenAlex
- W4396758709
- Cited by
- 115
- References
- 20
- Citations by year
- 2026: 62 · 2025: 48 · 2024: 5
- FWCI
- 43.13
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- International World Wide Web Conference (WWW) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion, cc-by)
- Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Funders
- National University of Singapore (A-8000969-00-00)
- Hong Kong University of Science and Technology
- Topics
- Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
- Social Sciences › Decision Sciences › Management Science and Operations Research › Forecasting Techniques and Applications
- Social Sciences › Transportation › Traffic Prediction and Management Techniques
- Keywords
- multivariate time series, domain identification, masking, zero-shot transfer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.