Reverso: Efficient Time Series Foundation Models for Zero-shot Forecasting

2026-02-19

AI Summary

This work proposes Reverso, a lightweight hybrid architecture for time series forecasting that combines long convolutions with linear recurrent neural networks (e.g., DeltaNet), complemented by tailored data augmentation and zero-shot inference optimization strategies. While existing foundation models achieve strong performance, they rely heavily on large Transformer architectures, resulting in excessive parameter counts and computational costs that hinder efficient deployment. In contrast, Reverso eschews conventional large-scale Transformers entirely, reducing model size by over two orders of magnitude while matching or even surpassing the zero-shot forecasting performance of current baselines. This approach substantially advances the Pareto frontier between efficiency and accuracy in time series modeling.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Natural Language Processing: (Large) Language Models
• Computer Vision: Large Vision Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Large language models for search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Learning time series foundation models has been shown to be a promising approach for zero-shot time series forecasting across diverse time series domains. Insofar as scaling has been a critical driver of performance of foundation models in other modalities such as language and vision, much recent work on time series foundation modeling has focused on scaling. This has resulted in time series foundation models with hundreds of millions of parameters that are, while performant, inefficient and expensive to use in practice. This paper describes a simple recipe for learning efficient foundation models for zero-shot time series forecasting that are orders of magnitude smaller. We show that large-scale transformers are not necessary: small hybrid models that interleave long convolution and linear RNN layers (in particular DeltaNet layers) can match the performance of larger transformer-based models while being more than a hundred times smaller. We also describe several data augmentation and inference strategies that further improve performance. This recipe results in Reverso, a family of efficient time series foundation models for zero-shot forecasting that significantly push the performance-efficiency Pareto frontier.

Problem

• time series foundation models
• zero-shot forecasting
• model efficiency
• parameter efficiency
• large-scale transformers

Innovation

• efficient foundation models
• zero-shot forecasting
• hybrid architecture
• DeltaNet
• time series