Live intelligence

Latest AI updates

3 updates
Research

SWIFT: Sample-Wise Instance Fine-Tuning of Time-Series Foundation Models for Forecasting

SWIFT is a new adaptation framework for time-series foundation models that fine-tunes on a per-sample or per-instance basis rather than requiring an additional target dataset. It leverages structural priors from the target instance's own observed history through Self-Retrieval and Frequency-Aware Loss, while using Deviation-Aware Dynamic Fusion to balance adaptation with general knowledge preservation. The authors report experiments across four TSFM backbones and Chronos Benchmark II subsets, with average MASE and WQL reductions of 6.41% and 4.83% across 24 backbone-dataset combinations. They position SWIFT as the first framework to investigate sample-wise instance fine-tuning of time-series foundation models for forecasting.

Time-series foundation modelsFine-tuning and adaptationForecasting benchmarks
Research

Study Stress-Tests Time-Series Foundation Models for Sampling Frequency Invariance

A new arXiv paper evaluates four prominent time-series foundation models—MOMENT, Chronos-2, Moirai-MoE, and TimesFM—under changes in sensor sampling frequency. It introduces three metrics, FSI, FDI, and CDI, to quantify embedding shifts, forecasting degradation, and downstream performance degradation. The study examines whether model behavior tracks lost information bandwidth rather than the sampling frequency value itself. It characterizes how invariance depends on the margin between native sampling frequency and task-required bandwidth, and how much TSFM choice and information loss explain performance changes.

Time-Series Foundation ModelsSampling Frequency InvarianceSensor Data
Research

FARE: Factor-Aware Retrieval with Error-Guided Fusion for Retail Sales Forecasting

FARE is a retrieval-augmentation method for short-term retail demand forecasting, where target weeks often combine promotion and calendar factors that have few exact historical precedents. It ranks historical weeks separately for each known target-week factor, merges the top matches into a factor reference under a fixed budget, and blends that with a shape reference drawn from similar sales histories. Two gating mechanisms, driven by shrunk errors from completed development weeks and by current disagreement among candidates, set the mix and control how strongly it corrects a backbone forecast; both gates stay fixed at test time. Across Favorita, M5, Rossmann and Rohlik, FARE cut WAPE of five backbones by 11.46%, 4.06%, 14.63% and 10.16% on average, and also beat supervised residual correction and retrieval augmentation with Chronos-2 frozen. Code is released publicly.

Retail demand forecastingRetrieval-augmented predictionTime-series foundation models