New LSTM stability method outperforms existing models

Researchers have developed a new method to ensure the stability of Long Short-Term Memory (LSTM) networks used in system identification, particularly for nonlinear dynamical systems like thermal processes. Their approach derives a sufficient condition for Input-to-State Stability (ISS) based on the infinity-norm, which relies on fewer network parameters than previous methods. This technique was validated on a thermal system, where the ISS-promoted LSTM outperformed both physics-based models and other recurrent neural network variants. AI

IMPACT Enhances the reliability of recurrent neural networks for modeling complex dynamic systems.

RANK_REASON The cluster contains an academic paper detailing a new method for improving LSTM network stability. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

• Gated Recurrent Unit
• Input-to-State Stability
• Recurrent Neural Networks
• Stefano De Carli

• paper
• other

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