KairosHope: A Next-Generation Time-Series Foundation Model for Specialized Classification via Dual-Memory Architecture

2026-05-18

AI Summary

This work addresses the limitations of existing time series foundation models, which suffer from the computational overhead of standard attention mechanisms and the neglect of classical statistical features. To overcome these issues, we propose KairosHope, a novel architecture that replaces conventional attention with a dual-memory design—comprising Titans for dynamic short-term memory and CMS for abstracting long-term historical patterns. A hybrid decision head integrates deep representations with time series statistical features extracted via tsfeatures. The model is trained using masked modeling and InfoNCE contrastive learning, and evaluated under the LP-FT transfer protocol. Extensive experiments on the UCR benchmark demonstrate that KairosHope significantly outperforms state-of-the-art methods, achieving particularly strong performance on highly causal time series classification tasks such as human activity recognition (HAR) and sensor-based applications.

Abstract

Time Series Foundation Models (TSFMs) have demonstrated notable success in general-purpose forecasting tasks; however, their adaptation to specialized classification problems remains constrained by the computational bottleneck of standard attention and the systematic omission of classical statistical knowledge. This technical report introduces KairosHope, a next-generation TSFM designed to reconcile massive generalization with analytical precision in classification tasks. The core of the proposal is the HOPE block, an architecture that replaces quadratic attention with a dual-memory system: Titans modules for dynamic short-term retention and a Continuum Memory System (CMS) for the abstraction of long-term historical context. To enrich the inductive bias, a Hybrid Decision Head is introduced, which fuses deep latent representations with deterministic statistical features extracted via tsfeatures package. KairosHope undergoes self-supervised pre-training on the massive Monash archive, combining Masked Time Series Modeling (MTSM) and contrastive learning (InfoNCE). Its subsequent adaptation to the UCR benchmark datasets is conducted through a rigorous Linear Probing and Full Fine-Tuning (LP-FT) protocol to prevent catastrophic forgetting. Empirical results demonstrate superior performance in domains characterized by strict temporal causality such as HAR or Sensor data. Consequently, KairosHope establishes a robust and efficient framework for the adaptation of foundation models to time series analysis.

Problem

• Time Series Foundation Models
• Specialized Classification
• Attention Bottleneck
• Statistical Knowledge Integration
• Temporal Causality

Innovation

• Dual-Memory Architecture
• Time Series Foundation Model
• HOPE Block
• Hybrid Decision Head
• Linear Probing and Full Fine-Tuning