TIDES model enhances irregular time series handling in selective SSMs

Researchers have introduced TIDES, a novel variant of selective state space models designed to improve handling of irregular time series data. Unlike existing selective SSMs like Mamba, which learn input-dependent time steps that deviate from physical time, TIDES maintains a consistent time step, allowing for native handling of irregular timestamps. This approach preserves the per-token expressivity of selective models while enabling better performance on benchmarks such as time series classification and ODE regression, particularly for datasets with irregular timestamps. AI

IMPACT Enhances capabilities for processing irregular time series data, potentially improving performance in fields like astronomy and climate science.

RANK_REASON New research paper detailing a novel model architecture. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Mamba
• Physiome
• Rui Manuel Barreira
• Selective state space models
• TIDES
• University of East Anglia

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →