# TS-Memory: Plug-and-Play Memory for Time Series Foundation Models

arXiv: https://arxiv.org/abs/2602.11550 · HTML: https://mlsanity.com/paper/2602.11550/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 21 references that are in the map.
- First posted on arXiv 12 Feb 2026.

## Builds on (most cited first, 12 of 12)

- [Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks](https://mlsanity.com/paper/1703.07015/) (2017)
- [One Fits All:Power General Time Series Analysis by Pretrained LM](https://mlsanity.com/paper/2302.11939/) (2023)
- [A Survey on RAG Meeting LLMs: Towards Retrieval-Augmented Large Language Models](https://mlsanity.com/paper/2405.06211/) (2024)
- [Transformer Memory as a Differentiable Search Index](https://mlsanity.com/paper/2202.06991/) (2022)
- [Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series](https://mlsanity.com/paper/2401.03955/) (2024)
- [RAGCache: Efficient Knowledge Caching for Retrieval-Augmented Generation](https://mlsanity.com/paper/2404.12457/) (2024)
- [OneNet: Enhancing Time Series Forecasting Models under Concept Drift by Online Ensembling](https://mlsanity.com/paper/2309.12659/) (2023)
- [Retrieval-Augmented Diffusion Models for Time Series Forecasting](https://mlsanity.com/paper/2410.18712/) (2024)
- [Proactive Model Adaptation Against Concept Drift for Online Time Series Forecasting](https://mlsanity.com/paper/2412.08435/) (2024)
- [Low-Rank Adaptation of Time Series Foundational Models for Out-of-Domain Modality Forecasting](https://mlsanity.com/paper/2405.10216/) (2024)
- [Decision-Focused Fine-Tuning of Time Series Foundation Models for Dispatchable Feeder Optimization](https://mlsanity.com/paper/2503.01936/) (2025)
- [Applying k-nearest neighbors to time series forecasting : two new approaches](https://mlsanity.com/paper/2103.14200/) (2021)

## Built on by (most cited first, 5 of 5)

- [Post-Training in Time Series Foundation Models: A Unifying Framework](https://mlsanity.com/paper/2607.20002/) (2026)
- [When Do Corrective Features Help? An Agent for Corrective Feature Discovery on Black-Box Forecasters](https://mlsanity.com/paper/2608.05207/) (2026)
- [When Does Retrieval Help Time-Series Forecasting?](https://mlsanity.com/paper/2609.20193/) (2026)
- [When Tomorrow Becomes Today: Self-Evolving Policies for Agentic Time-Series Forecasting](https://mlsanity.com/paper/2609.24862/) (2026)
- [Instance-Adaptive Prompts as Context for Time-Series Foundation Models](https://mlsanity.com/paper/2609.34786/) (2026)

## Abstract (arXiv)

Time Series Foundation Models (TSFMs) achieve strong zero-shot forecasting through large-scale pre-training, but adapting them to downstream domains under distribution shift remains challenging. Existing solutions face a trade-off: Parametric Adaptation can cause catastrophic forgetting and requires costly multi-domain maintenance, while Non-Parametric Retrieval improves forecasts but incurs high inference latency due to datastore search. We propose Parametric Memory Distillation and implement it as TS-Memory, a lightweight memory adapter that augments frozen TSFMs. TS-Memory is trained in two stages. First, we construct an offline, retrieval-leakage-safe kNN teacher that synthesizes confidence-aware quantile targets from retrieved futures. Second, we distill this retrieval-induced distributional correction into a lightweight memory adapter via confidence-gated supervision. During inference, TS-Memory fuses memory and backbone predictions with constant-time overhead, enabling retrieval-free deployment. Experiments across diverse TSFMs and benchmarks demonstrate consistent improvements in both point and probabilistic forecasting over representative adaptation methods, with efficiency comparable to the frozen backbone. Code: https://github.com/sisuolv/TS-Memory.