Lattice Random Walk Discretisations of Stochastic Differential Equations
By: Samuel Duffield , Maxwell Aifer , Denis Melanson and more
Potential Business Impact:
Makes computers understand complex math faster.
We introduce a lattice random walk discretisation scheme for stochastic differential equations (SDEs) that samples binary or ternary increments at each step, suppressing complex drift and diffusion computations to simple 1 or 2 bit random values. This approach is a significant departure from traditional floating point discretisations and offers several advantages; including compatibility with stochastic computing architectures that avoid floating-point arithmetic in place of directly manipulating the underlying probability distribution of a bitstream, elimination of Gaussian sampling requirements, robustness to quantisation errors, and handling of non-Lipschitz drifts. We prove weak convergence and demonstrate the advantages through experiments on various SDEs, including state-of-the-art diffusion models.
Similar Papers
An Explicit Euler-type Scheme for Lévy-driven SDEs with Superlinear and Time-Irregular Coefficients
Numerical Analysis
Makes computer models of messy real-world events more accurate.
Adaptive Stochastic Coefficients for Accelerating Diffusion Sampling
CV and Pattern Recognition
Makes AI create pictures faster and better.
Filtering and 1/3 Power Law for Optimal Time Discretisation in Numerical Integration of Stochastic Differential Equations
Systems and Control
Improves computer predictions of moving things.