Researchers have introduced TIDES, a novel selective state space model (SSM) designed to effectively handle irregular time series data. Unlike existing models like Mamba, which learn a step size that deviates from physical time, TIDES maintains a step size equivalent to the physical time gap. This is achieved by shifting input dependence to the diagonal state matrix, allowing for native handling of irregular timestamps without sacrificing per-token expressivity. TIDES has demonstrated strong performance on benchmarks such as the Fading Flash experimental benchmark, the UEA time series classification, and the Physiome ODE regression, outperforming current state-of-the-art architectures on natively irregular datasets. AI
IMPACT Enhances the capability of AI models to process and analyze time-series data with irregular timestamps, potentially improving applications in fields like astronomy and climate science.
RANK_REASON The cluster describes a new academic paper introducing a novel model architecture. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
