Search papers, labs, and topics across Lattice.
This paper provides a non-asymptotic error analysis for Sequential Monte Carlo (SMC) methods that utilize biased proposals in the context of conditional diffusion sampling. By decomposing total error into kernel bias and finite-particle Monte Carlo error, the authors extend existing theoretical frameworks to offer a principled control over these biases. The key result is the derivation of the first non-asymptotic error bound that simultaneously addresses multiple sources of error in score-based diffusion models, enhancing the reliability of generative sampling techniques.
Non-asymptotic error bounds reveal that biased proposals in SMC can be effectively managed, significantly improving the reliability of conditional diffusion sampling.
Sequential Monte Carlo (SMC) methods are a natural tool for post-hoc conditioning of pretrained generative models, but in many applications the mutation kernels used by the particle system are biased approximations of an ideal Feynman--Kac flow. This paper develops a non-asymptotic error analysis for such SMC samplers. Under forward-smoothing forgetting conditions, we decompose the total error into a kernel bias, measuring the effect of replacing the ideal transition kernels by approximate ones, and a finite-particle Monte Carlo error. Our approach relies on extending local Doeblin-type conditions and Lyapunov drift arguments for Markov kernels to conditional distributions, thereby enabling a principled control of the bias. We then instantiate this general framework for conditional sampling with score-based diffusion models, and derive the first non-asymptotic error bound that jointly controls initialization error, time discretization, and score approximation in the reverse diffusion dynamics as well as finite-particle Monte Carlo error.
