Data-driven Optimal Sensor Placement for Aerosol Monitoring in the UK

Key points

• A trained convolutional Gaussian neural process is embedded within the DeepSensor optimisation framework.1
• Genetic optimisation with lazy evaluation and caching reduces mean computational time by 60% (10$\pm$3 versus 25$\pm$7 minutes) while producing sequential placement performance comparable to the greedy solver.1
• Using the CAMS reference field to construct a retrospective oracle benchmark, Joint NLL improves by about 22% for Joint-MI-driven genetic placement after five locations for aerosol stations are identified.1
• These results show a computationally efficient approach for data-informed expansion of environmental observing systems.1

Sources (1)

• [1]Data-driven Optimal Sensor Placement for Aerosol Monitoring in the UK
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 07:25 PM

  We develop a data-driven sensor-placement framework and apply it to aerosol monitoring in the United Kingdom (UK).

  A trained convolutional Gaussian neural process is embedded within the DeepSensor optimisation framework.

Extractive summary: sentences quoted from the sources.