Embedded Files
Conference Poster_v2.pdfA Dynamic Network and Representation Learning Approach for Quantifying Economic Growth from Satellite Imagery
A Dynamic Network and Representation Learning Approach for Quantifying Economic Growth from Satellite Imagery
Accepted and presented at 2018 NeurIPS workshop: Machine Learning for Developing World.
arXiv link:
1. Propose a dynamic unsupervised gravitational network drawn from nightlight satellite imagery from developing countries.
2. Mining the network for representation learning for dynamic changing feature embeddings.
3. Quantify the economic growth by doing spatiotemporal economic indicator prediction.
Puerto Rico damaged by Hurricane Maria in 2017
Network for Urban Development and Human Behavior Monitoring from Satellite Imagery
Network for Urban Development and Human Behavior Monitoring from Satellite Imagery
1. Propose a dynamic unsupervised gravitational network drawn from nightlight satellite imagery.
2. Monitoring the urban development trend as well as human behavior by analysis the network dynamic attributes.
Dynamic Network on Texas from 2012 to 2018
Dynamic Network on Texas from 2012 to 2018
Google Sites
Report abuse