微分可能なリサンプリングに向けて
Towards Differentiable Resampling
粒子フィルタのリサンプリングを学習型ニューラルネット（particle transformer）で置き換え、微分可能な粒子フィルタとして end-to-end 学習を可能にした研究。
著者: Michael Zhu, Kevin Murphy, Rico Jonschkowski
分類: cs.LG, cs.RO, stat.ML
原文アブストラクト
Resampling is a key component of sample-based recursive state estimation in particle filters. Recent work explores differentiable particle filters for end-to-end learning. However, resampling remains a challenge in these works, as it is inherently non-differentiable. We address this challenge by replacing traditional resampling with a learned neural network resampler. We present a novel network architecture, the particle transformer, and train it for particle resampling using a likelihood-based loss function over sets of particles. Incorporated into a differentiable particle filter, our model can be end-to-end optimized jointly with the other particle filter components via gradient descent. Our results show that our learned resampler outperforms traditional resampling techniques on synthetic data and in a simulated robot localization task.
