Search papers, labs, and topics across Lattice.
TimEE is a novel 4.5M-parameter foundation model designed for end-to-end time series classification (TSC) using in-context learning, eliminating the need for per-dataset training. By leveraging a meta-training approach on synthetic TSC tasks, TimEE outputs class distributions directly from labeled support sets and query time series in a single forward pass. Remarkably, despite being trained exclusively on synthetic data, TimEE achieves state-of-the-art performance on the UCR benchmark, ranking first in ROC AUC and third in accuracy among various methods, including foundation models and supervised deep learning approaches.
Achieving state-of-the-art TSC performance without real data, TimEE redefines the potential of synthetic pre-training in classification tasks.
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. While effective, this decoupling optimizes representation learning independently of the classification objective, requires per-dataset training, and prevents the model from exploiting label information during inference. We introduce TimEE, a 4.5M-parameter foundation model for end-to-end TSC via in-context learning. Given a labeled support set and a query time series, TimEE directly outputs a predicted class distribution in a single forward pass with no per-dataset training required. Following the prior-data fitted network (PFN) framework, TimEE is meta-trained exclusively on synthetic TSC tasks, where each task contains time series with distinct class identities arising from structured distributional shifts in the generative process. Despite seeing no real time series during pre-training, TimEE ranks first in ROC AUC (and third on accuracy) on the UCR benchmark among all compared methods, which include both foundation models and supervised deep learning baselines. To our knowledge, TimEE is the first purely synthetic-pretrained model to reach state-of-the-art performance on the UCR benchmark. These results establish end-to-end ICL with synthetic priors as a compelling, largely unexplored direction for TSC, with scaling, prior design, and richer generation mechanisms as natural avenues for improvement. Code is publicly available at http://github.com/automl/timee.
