PulseAugur
EN
LIVE 10:51:08

New methods for uncertainty propagation in random neural networks developed

By PulseAugur Editorial · [2 sources] ·

Researchers have developed new analytical and particle-based methods for uncertainty propagation in random neural network models. These methods leverage the piecewise-linear structure of the Leaky ReLU activation function to approximate the neural network's output, enabling the computation of analytical expressions for its probability density and characteristic functions. The framework has been extended to autoregressive models representing dynamical systems, with numerical experiments on the Lorenz-63 system and Kuramoto-Sivashinsky equation demonstrating accurate uncertainty propagation. AI

IMPACT These methods could improve the reliability and interpretability of neural networks in complex dynamical systems.

RANK_REASON The cluster contains a research paper detailing new methods for uncertainty propagation in neural networks.

Read on arXiv cs.NE (Neural & Evolutionary) →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New methods for uncertainty propagation in random neural networks developed

COVERAGE [2]

  1. arXiv cs.NE (Neural & Evolutionary) TIER_1 English(EN) · Daniele Venturi ·

    Uncertainty propagation in auto-regressive random neural network models

    We develop analytical and particle-based methods for uncertainty propagation in random neural network models, where both the inputs and network parameters are allowed to be random. Building on the piecewise-linear structure of the Leaky ReLU activation function, we derive a local…

  2. arXiv stat.ML TIER_1 English(EN) · Janice Adams, Daniele Venturi ·

    Uncertainty propagation in auto-regressive random neural network models

    arXiv:2608.20483v1 Announce Type: new Abstract: We develop analytical and particle-based methods for uncertainty propagation in random neural network models, where both the inputs and network parameters are allowed to be random. Building on the piecewise-linear structure of the L…

RELATED ENTITIES

RELATED TOPICS