Causal Discovery of Dynamic Models for Predicting Human Spatial Interactions

Abstract

Identifying the main features and learning the causal relationships of a dynamic system from timeseries of sensor data are key problems in many real-world robot applications. In this paper, we propose an extension of a state-of-the-art causal discovery method, PCMCI, embedding an additional feature-selection module based on transfer entropy. Starting from a prefixed set of variables, the new algorithm reconstructs the causal model of the observed system by considering only its main features and neglecting those deemed unnecessary for understanding the evolution of the system. We first validate the method on a toy problem and on synthetic data of brain network, for which the ground-truth models are available, and then on a real-world robotics scenario using a large-scale time-series dataset of human trajectories. The experiments demonstrate that our solution outperforms the previous state-of-the-art technique in terms of accuracy and computational efficiency, allowing better and faster causal discovery of meaningful models from robot sensor data.

Continual Deep Learning for Time Series Modeling

“…Currently, popular causal discovery methods utilize only static data or pre-processed time series data in advance, making them not very suitable for realworld robotics cases. Continual learning may address this limitation in the causal discovery methods [132] but is under-investigated for robot application [133]. Castri et al [133] focused on the constraint-based methods for causal discovery and outlined their approach of re-learning the causal model during observed scenario changes and during a new set of interventions.…”

“…Continual learning may address this limitation in the causal discovery methods [132] but is under-investigated for robot application [133]. Castri et al [133] focused on the constraint-based methods for causal discovery and outlined their approach of re-learning the causal model during observed scenario changes and during a new set of interventions. The new inference matric of the causal model is checked against the matric of the old causal model to discover the still valid causal links from the old model for the new model.…”