Nonparametric inference for density-dependent McKean--Vlasov diffusions
The episode discusses a paper titled "Nonparametric inference for density-dependent McKean–Vlasov diffusions." The authors propose a method that simplifies complex, high-dimensional systems into one-dimensional problems. By using a ReQU neural network sieve and an endpoint-adapted graded approximation, they achieve highly accurate estimation of unknown functions with strong, fast convergence rates.