arXiv:2606. 09861v1 Announce Type: cross Abstract: While Next-Token Prediction (NTP) has unified LLM pretraining, its adaptation to unbounded, continuous time series (TS) remains open.
By Yunhao Zhang, Ruiying Qi, Jiale Zheng, Jianfeng Zhang, Lujia Pan, Junchi Yan
The paper investigates whether the initial recurrent state in time‑series foundation models (TSFMs) can be leveraged to improve forecasting accuracy. Using the xLSTM‑based zero‑shot forecaster TiRex, the authors freeze the backbone and train a small head that writes a context vector into the initial state, demonstrating modest but consistent gains across several datasets. The study shows that the initial state can serve both as a per‑dataset adaptation surface and as a channel for exogenous information, suggesting a new avenue for enhancing recurrent TSFMs.
By Bernhard Voggenberger, Christian Ganh\"or, Levente Z\'olyomi, Sebastian B\"ock, Sepp Hochreiter
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:2601. 16632v4 Announce Type: replace-cross Abstract: Time series forecasting has witnessed significant progress with deep learning.
By Haonan Yang, Jianchao Tang, Zhuo Li
arXiv:2607. 07500v1 Announce Type: cross Abstract: Time series classification (TSC) is dominated by a two-stage paradigm: train a feature encoder -- either from scratch on the target dataset or via pretraining on large corpora -- and then fit a task-specific classifier on top.
By Jaris K\"uken, Shi Bin Hoo, Martin Mr\'az, Frank Hutter, Lennart Purucker
The paper introduces PaCTS, a method that generates instance‑adaptive latent prompts—continuous embedding tokens—to provide compact contextual information for frozen time‑series foundation models (TSFMs). These prompts are constructed from instance‑specific global statistics and refined with segment‑level temporal data, enabling the model to capture both global characteristics and local temporal variations. Experiments show that PaCTS improves forecasting performance across various context lengths and model architectures, often outperforming the same backbone with double the context while reducing inference computation, and it also offers stronger improvements and better out‑of‑distribution generalization compared to weight‑space adaptation methods.
arXiv:2609.15344v1 Announce Type: new
Abstract: We study the adaptation of pretrained language models to univariate time-series forecasting through a parameter-efficient transfer learning framework,...
By Tamanna Kumavat, Georg Brunner, Kyriakos Flouris
AdaCast introduces a conditional parameter generation framework that tailors time‑series forecasting models to each input by producing low‑rank updates for a frozen pretrained foundation model. Unlike static adaptation methods that apply a single set of updates to all inputs, AdaCast generates input‑specific updates during both training and inference. Experiments on six public benchmarks show that AdaCast consistently outperforms static baselines in in‑domain forecasting and improves zero‑shot generalization to unseen datasets across domains.
By Darahaas Nallagatla, Darryl Cherian Jacob, Pan He
arXiv:2611. 05993v1 Announce Type: new Abstract: Time series forecasting is critical for demand prediction and resource management in modern Web services, yet applying pretrained Time Series Foundation Models (TSFMs) to downstream tasks remains challenging due to domain shifts.
By Yuqian Wu, Qingxiang Liu, Yuxin Guo, Weifeng Su, Jian Li, Pengfei Xu, Xi Chen, Yutao Yue, Yuxuan Liang
The paper introduces iAmTime, a time‑series foundation model that uses instruction‑conditioned in‑context learning to adapt to tasks at inference time. iAmTime represents each episode as a structured prompt with semantic tokens that focus on specific time‑series regions, enabling the model to infer task structure from input‑output demonstrations. Trained on large real and synthetic corpora across forecasting, imputation, reconstruction, classification, anomaly detection, and source de‑mixing, iAmTime outperforms strong baselines on zero‑shot probabilistic and point forecasting while matching or exceeding performance on several non‑forecasting tasks.
By Anish Saha, Konstantin Shmakov
Dual-Context Analog Retrieval for Time Series Forecasting (DuoTS) introduces a two-stage forecasting approach that first generates a base forecast and then refines it patch by patch. Each refinement step fuses a current context, which focuses on recent tokens, with a detail context that incorporates retrieved analogs and their subsequent trajectories. Experiments on real-world datasets demonstrate that DuoTS achieves state‑of‑the‑art performance, and ablation studies confirm the importance of both contexts and the refinement mechanism.
By Jung Min Choi, Ngoc Son Le, Ibram Abdelmalak, Vijaya Krishna Yalavarthi, Lars Schmidt-Thieme
arXiv:2602. 00620v2 Announce Type: replace-cross Abstract: The zero-shot evaluation of time series foundation models (TSFMs) for classification typically uses a frozen encoder followed by a task-specific classifier.
By Juntao Fang, Shifeng Xie, Shengbin Nie, Yuhui Ling, Yuming Liu, Zijian Li, Keli Zhang, Lujia Pan, Themis Palpanas, Ruichu Cai