world2vec：通过潜在动力学学习构建机制性世界模型
world2vec: Mechanistic world models through latent dynamics learning
Tobias Schmidt, Rodrigo Gonzalez Laiz, Krisztián Szuppinger, Steffen Schneider
AI 速览：world2vec 仅从观测与动作中学习可长期忠实复现真实系统的潜在动力学，以更少参数实现控制与系统辨识。
World models can achieve strong control performance without faithfully capturing the dynamics of their environment, limiting their potential for mechanistic interpretability and scientific discovery. Here, we introduce world2vec to learn mechanistic world models without reward information directly from observations and actions. World2vec learns latent dynamics that faithfully reproduce the true system over long prediction horizons, with at least one order of magnitude fewer parameters than DreamerV3, TD-MPC2 and LeWM. This enables state of the art performance in controlling continuous environments from pixels and accurate system identification across a benchmark of 71 attractors, demonstrating how world models are amenable to scientific discovery.