DireSMC: Steering Diffusion Models to Rare Events
What it is
This paper proposes DireSMC, a sequential Monte Carlo method that guides weighted samples in diffusion models toward rare events. It uses an analytical relaxation of the event set to provide calibrated probability estimates and efficient sampling for events with probabilities as low as 10^-5.
The problem it solves
Standard Monte Carlo estimation of rare event probabilities in diffusion models is computationally intractable because the required sample size scales inversely with the event probability (1/p0[E]).
What you could use it for
- Estimating probabilities of extreme weather events using climate emulators
- Identifying rare molecular configurations in molecular dynamics simulations
- Sampling rare states in materials design workflows
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.