Environmental sensors are crucial for monitoring weather conditions and the impacts of climate change. However, it is challenging to maximise measurement informativeness and place sensors efficiently, particularly in remote regions like Antarctica. Probabilistic machine learning models can evaluate placement informativeness by predicting the uncertainty reduction provided by a new sensor. Gaussian process (GP) models are widely used for this purpose, but they struggle with capturing complex non-stationary behaviour and scaling to large datasets. This paper proposes using a convolutional Gaussian neural process (ConvGNP) to address these issues. A ConvGNP uses neural networks to parameterise a joint Gaussian distribution at arbitrary target locations, enabling flexibility and scalability. Using simulated surface air temperature anomaly over Antarctica as ground truth, the ConvGNP learns spatial and seasonal non-stationarities, outperforming a non-stationary GP baseline. In a simulated sensor placement experiment, the ConvGNP better predicts the performance boost obtained from new observations than GP baselines, leading to more informative sensor placements. We contrast our approach with physics-based sensor placement methods and propose future work towards an operational sensor placement recommendation system. This system could help to realise environmental digital twins that actively direct measurement sampling to improve the digital representation of reality.

翻译：环境传感器对于监测天气状况和气候变化影响至关重要。然而，如何最大化测量信息量并高效部署传感器仍面临挑战，尤其是在南极等偏远地区。概率机器学习模型可通过预测新传感器带来的不确定性降低程度来评估部署信息量。高斯过程模型虽被广泛用于此目的，但在捕捉复杂非平稳行为和大规模数据扩展方面存在困难。本文提出采用卷积高斯神经过程解决上述问题。该模型通过神经网络在任意目标位置参数化联合高斯分布，兼具灵活性与可扩展性。以模拟的南极地表气温异常为基准，ConvGNP学习了空间与季节非平稳特征，其性能优于非平稳高斯过程基线模型。在模拟传感器布局实验中，ConvGNP能比高斯过程基线模型更准确地预测新观测值带来的性能提升，从而指导更具信息量的传感器部署。我们将该方法与基于物理的传感器布局方案进行对比，并提出构建可操作型传感器布局推荐系统的未来研究方向。该系统将有助于实现主动引导测量采样以优化数字孪生环境表征的环境数字孪生系统。