{"slug": "quantflow-a-federated-mamba-based-post-transformer-foundation-model-for-time", "title": "QuantFlow: A Federated Mamba-Based Post-Transformer Foundation Model for Time-Series Forecasting", "summary": "Researchers introduced QuantFlow, a federated Mamba-based foundation model for time-series forecasting that combines inverted sequence embedding, bidirectional state-space decoders, quantile regression, and federated learning. The model achieved mean squared errors of 0.2834 on ETTm1 and 0.2218 on Weather, and maintained accuracy in a 20-client non-IID deployment after three communication rounds without centralizing raw data. QuantFlow demonstrates selective state-space modeling as a scalable, uncertainty-aware, and privacy-conscious approach for time-series prediction.", "body_md": "arXiv:2607.02632v1 Announce Type: new\nAbstract: Time-series forecasting supports decisions in finance, en-ergy, transportation, public health, and industrial monitoring. Recent foundation models improve transfer across forecast-ing tasks, but many depend on centralized data and Trans-former attention, which restricts their use for long, high-di-mensional, and privacy-sensitive signals. This paper presents QuantFlow, a probabilistic forecasting framework that com-bines inverted sequence embedding, bidirectional Mamba state-space decoders, quantile regression, and federated learning. Each variable is embedded over the complete ob-servation window, processed in forward and reverse direc-tions, and projected to five conditional quantiles. TSMixup expands temporal diversity through Dirichlet-weighted inter-polation while preserving sequence structure. Experiments cover cryptocurrency, traffic, electricity, Electricity Trans-former Temperature, influenza, and weather data. QuantFlow obtains mean squared errors of 0.2834 on ETTm1 and 0.2218 on Weather, and a 20-client non-IID deployment retains use-ful accuracy after three communication rounds without cen-tralizing raw records. The results indicate that selective state-space modelling is a promising basis for scalable, uncer-tainty-aware, and privacy-conscious time-series prediction, while also revealing limitations on irregular epidemiological signals and long-horizon generalization.", "url": "https://wpnews.pro/news/quantflow-a-federated-mamba-based-post-transformer-foundation-model-for-time", "canonical_source": "https://arxiv.org/abs/2607.02632", "published_at": "2026-07-07 04:00:00+00:00", "updated_at": "2026-07-07 04:11:04.829085+00:00", "lang": "en", "topics": ["machine-learning", "ai-research", "ai-infrastructure", "ai-ethics"], "entities": ["QuantFlow", "Mamba", "ETTm1", "Weather", "TSMixup", "Dirichlet"], "alternates": {"html": "https://wpnews.pro/news/quantflow-a-federated-mamba-based-post-transformer-foundation-model-for-time", "markdown": "https://wpnews.pro/news/quantflow-a-federated-mamba-based-post-transformer-foundation-model-for-time.md", "text": "https://wpnews.pro/news/quantflow-a-federated-mamba-based-post-transformer-foundation-model-for-time.txt", "jsonld": "https://wpnews.pro/news/quantflow-a-federated-mamba-based-post-transformer-foundation-model-for-time.jsonld"}}