Timer‑M1 is a pretrained multivariate time‑series foundation model that learns with primitives for zero‑shot forecasting. It uses a primitive‑based data synthesis and pretraining pipeline to generate series with shared temporal and relational patterns, assembling them into multivariate samples with distinct channel roles. The model employs gated two‑dimensional Transformer blocks that dynamically allocate cross‑variate attention, achieving top performance on several large‑scale forecasting benchmarks.
By Haoran Zhang, Haixuan Liu, Xingjian Su, Yong Liu, Zhi Chen, Yuxuan Wang, Jianmin Wang, Mingsheng Long
arXiv:2602. 02288v3 Announce Type: replace Abstract: Current time-series forecasting models are primarily based on transformer-style neural networks.
By Zheng Li, Jerry Cheng, Huanying Gu
arXiv:2608. 06223v1 Announce Type: new Abstract: While deep learning models, particularly transformer-based architectures, have shown impressive performance in time series forecasting, the application of retrieval-augmented generation (RAG) in this domain remains limited.
By Yixiong Xiao, Congxi Xiao, Jingbo Zhou
arXiv:2607. 23146v1 Announce Type: new Abstract: Inspired by recent breakthroughs in large language models for natural language processing, foundation models have emerged as a promising paradigm for zero-shot time series forecasting, enabling accurate predictions on datasets never seen during pre-training.
By Morad Laglil, Bertrand Pracca, Emilie Devijver, Eric Gaussier
arXiv:2608. 14054v1 Announce Type: new Abstract: Time series forecasting with pretrained foundation models has demonstrated strong zero-shot capabilities.
By Juan Pablo Villa Serna, Rohan Asthana, Vasileios Belagiannis
While deep learning models, particularly transformer-based architectures, have shown impressive performance in time series forecasting, the application of retrieval-augmented generation (RAG) in this domain remains limited. Since RAG has proven effective in enhancing the capabilities of large language models by incorporating relevant external information, retrieving similar time series sequences as references might also improve accuracy in time series forecasting tasks.
NeST is a framework that adapts large language models (LLMs) for continuous time‑series forecasting by creating neighborhood‑aware text prototypes and aligning them with temporal representations through a nearest‑neighbor contrastive objective. It retrieves the most relevant prototypes and uses them to conditionally modulate time‑series features, enabling more effective integration of textual and temporal information. Experiments show that NeST outperforms state‑of‑the‑art methods on eight benchmarks, reduces MSE by 1.2% for long‑term forecasting, improves zero‑shot forecasting by 4.9%, and boosts R² by 3.3% on a real‑world photovoltaic power forecasting task.
By Jayanie Bogahawatte, Sachith Seneviratne, Maneesha Perera, Saman Halgamuge
The paper introduces a simple, model‑agnostic time‑domain augmentation called Sliding‑Window Reordering with Overlap Averaging. It transforms the joint input‑target sequence into overlapping windows, randomly reorders a fraction of them based on a variance criterion, and reconstructs the sequence by averaging overlaps to generate synthetic samples with controlled variation and minimal temporal distortion. Experiments show strong performance gains across nine long‑term forecasting benchmarks and four short‑term traffic benchmarks, with detailed ablations and diagnostics highlighting the effectiveness of each design choice.
By Jafar Bakhshaliyev, Johannes Burchert, Niels Landwehr, Lars Schmidt-Thieme
Aurora‑X is a billion‑parameter time‑series foundation model designed for extreme forecasting tasks. It employs a progressive curriculum that starts with channel‑independent pretraining, then adds cross‑variable dependencies, variable context and horizon lengths, and optional future covariates during mid‑training. A variable‑resolution post‑training stage allows adjustable temporal spans per token at inference, while a pattern‑guided mixture‑of‑experts expands capacity through sparse activation and expert specialization. An implicit quantile network head predicts arbitrary quantiles, enhancing probabilistic forecasting flexibility. Experiments on GIFT‑Eval, TIME, FEV‑Bench, TFB, and DAG‑Bench show state‑of‑the‑art performance against both pretrained TSFMs and task‑specific supervised models.
By Xingjian Wu, Chenjuan Guo, Xiangfei Qiu, Zhigang Hu, Hanyin Cheng, Peng Chen, Yang Shu, Jilin Hu, Bin Yang
arXiv:2505. 15354v3 Announce Type: replace Abstract: Time-series forecasting is a critical task in various business domains, but it remains inherently challenging.
By Hamza Cherkaoui, Malik Tiomoko, Giuseppe Paolo, Zhang Yili, Yu Meng, Zhang Keli, Hafiz Tiomoko Ali
arXiv:2605. 00015v2 Announce Type: replace-cross Abstract: Time Series Foundation Models (TSFMs) have demonstrated strong generalization capability and data efficiency in time series forecasting through large-scale pretraining.
By Siyang Li, Yize Chen, Zijie Zhu, Yuxin Pan, Yan Guo, Ming Huang, Hui Xiong
SteerCast is a retrieval‑based latent steering technique that enhances decoder‑only time‑series forecasters during inference without updating model parameters. It builds a database from the training set, storing each history window’s representation and a steering vector that captures the difference between the model’s prediction and the ground‑truth continuation in latent space. During testing, SteerCast retrieves nearest neighbors, aggregates their steering vectors, and injects the resulting signal into the forecaster’s hidden states at each autoregressive step, guiding predictions toward trajectories seen in similar training cases. Experiments on diverse multivariate benchmarks and multiple horizons demonstrate consistent accuracy improvements over fine‑tuned backbones and other retrieval‑based baselines, all while requiring no additional training beyond the original fine‑tuning and using only the training set as a retrieval corpus.
By Van Dai Do, Huu Hiep Nguyen, Minh Hoang Nguyen, Hung Le