Track trials, findings, and clinical evidence as it lands.
Search papers, labs, and topics across Lattice.
This paper introduces Retrieval-to-Recognition (R2R), a plug-and-play post-training method for time series foundation models (TSFMs) that enhances classification by explicitly retrieving instance-level numerical knowledge. By decoupling a query encoder to extract discriminative patterns and fusing this explicit knowledge with TSFM embeddings, R2R enables joint knowledge transfer at both parameter and feature levels, improving the utilization of pre-trained knowledge. Experiments across 158 datasets demonstrate consistent performance gains for state-of-the-art TSFMs.
Explicitly retrieving instance-level numerical knowledge during post-training unlocks significant classification performance gains for existing time series foundation models.
Time series foundation models (TSFMs) have achieved notable progress in classification tasks through large-scale pretraining. However, their post-training practice mainly focuses on implicit parameter fine-tuning, where task-relevant information is entangled with abundant irrelevant knowledge in vast parameter spaces. This may hinder TSFMs from precisely activating useful knowledge for downstream classification. To this end, we propose Retrieval-to-Recognition (R2R), a plug-and-play post-training paradigm that adapts pretrained TSFMs to time series classification. Rather than pretraining a new foundation model, R2R improves how existing TSFMs are post-trained: it introduces a decoupled query encoder to extract discriminative patterns and retrieve instance-level numerical knowledge from a constructed knowledge base. The retrieved explicit knowledge is then fused with the TSFM's implicit parametric embeddings, enabling joint knowledge transfer at both the parameter and feature levels. In this way, R2R allows TSFMs to recall task-relevant patterns learned during pretraining, improving the efficiency of knowledge utilization in downstream classification transfer. Experiments on 158 time series datasets show that R2R consistently improves the classification performance of state-of-the-art TSFMs. Ablation and knowledge-enhanced studies further validate the effectiveness of the proposed framework.