Data for Development: Research
The Data for Development initiative enables and supports sustained interactions between experts in both data science and development.
Selected Current Projects
Landscape of brick manufacturing in Bangladesh: an application of deep learning for sustainable development
- Authors: Jihyeon Lee and Nina Brooks
- Use high resolution satellite imagery to map brick kilns across entire country
- Identify total number of kilns, percent of population exposed to pollution from kilns, and number of violations of various regulations
- For example, 13 percent of health facilities and 11 percent of schools are within one kilometer of a kiln
- Demonstrate higher pollution during brick season in proximity of kilns
- Paper in preparation for submission to PNAS
- Link: https://lubylab.stanford.edu/assessing-brick-kilns-number-location-and-…
Using remote sensing data for causal inference
- Author: Brandon de la Cuesta
- Combination of remote sensing data and revolution of machine learning approaches has generated huge quantities of data on developing countries
- New sources of high-frequency, high-resolution data
- Majority of existing work focuses on descriptive task (i.e., mapping poverty)
Data can also be used to help us answer core questions in development and political economy
- Does political competition actually lead to more public goods provision and better economic outcomes?
- How much do effects of climate change (e.g., drought, fires), contribute to political instability?
Goal: Investigate promises and perils of using datasets produced by ML algorithms to answer causal questions
- Evaluate whether prediction error is larger for vulnerable or marginalized populations
- Understand implication of and temporal bias in availability of ground truth data
- Develop optimal sampling algorithm to reduce correlation between prediction error magnitude and key covariates
Revisiting core findings in development literature with new data
- Author: Brandon de la Cuesta
- DDI researchers have produced first-ever grid-cell measure of asset wealth and consumption in Sub-Saharan Africa
- Seminal findings in political economy and development literature utilize poorly measured or highly aggregated measures of wealth
- Goal: Identify high-impact applications of wealth estimates produced via Convolutional Neural Net (CNN)
- Does democracy lead to better public goods provision? If so, where and by how much?
- What are the effects of ethnic inequality on key political and economic outcomes?
- How much do leaders favor co-ethnics in the distribution of government revenues?
Lee, J., Grosz, D., Uzkent, B., Zeng, S., Burke, M., Lobell, D. and Ermon, S., 2021, May. Predicting Livelihood Indicators from Community-Generated Street-Level Imagery. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 35, No. 1, pp. 268-276).
Lee, J., Brooks, N.R., Tajwar, F., Burke, M., Ermon, S., Lobell, D.B., Biswas, D. and Luby, S.P., 2021. Scalable deep learning to identify brick kilns and aid regulatory capacity. Proceedings of the National Academy of Sciences, 118(17).
Burke, M., Driscoll, A., Lobell, D.B. and Ermon, S., 2021. Using satellite imagery to understand and promote sustainable development. Science, 371(6535).
Ayush, K., Uzkent, B., Meng, C., Tanmay, K., Burke, M., Lobell, D. and Ermon, S., 2020. Geography-aware self-supervised learning. arXiv preprint arXiv:2011.09980.
Yeh, C., Perez, A., Driscoll, A., Azzari, G., Tang, Z., Lobell, D., Ermon, S. and Burke, M., 2020. Using publicly available satellite imagery and deep learning to understand economic well-being in Africa. Nature communications, 11(1), pp.1-11.
Ayush, K., Uzkent, B., Kumar T, Burke M, Lobell, D.and Ermon, S., 2021, May. Efficient Poverty Mapping from High Resolution Remote Sensing Images. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 35, No. 1, pp. 12-20).
Uzkent, B. and Ermon, S., 2020. Learning when and where to zoom with deep reinforcement learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 12345-12354).
Uzkent, B., Yeh, C. and Ermon, S., 2020. Efficient object detection in large images using deep reinforcement learning. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 1824-1833).
Sarukkai, V., Jain, A., Uzkent, B. and Ermon, S., 2020. Cloud removal from satellite images using spatiotemporal generator networks. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 1796-1805).
Aung HL, Uzkent B, Burke M, Lobell D, Ermon S. “Farmland Parcel Delineation Using Spatio-temporal Convolutional Networks." arXiv preprint arXiv:2004.05471. 2020 Apr 11. (IEEE CVPR 2020 Workshop on Agriculture-Vision: Challenges & Opportunities For Computer Vision in Agriculture)
Uzkent Burak, and Stefano Ermon. "Learning When and Where to Zoom with Deep Reinforcement Learning." arXiv preprint arXiv:2003.00425 (2020). (IEEE CVPR 2020)
*Ayush, Kumar, *Burak Uzkent, Marshall Burke, David Lobell, and Stefano Ermon. "Generating Interpretable Poverty Maps using Object Detection in Satellite Images." arXiv preprint arXiv:2002.01612 (2020). (*Shared First Author) (IJCAI 2020)
Uzkent Burak, Christopher Yeh, and Stefano Ermon. "Efficient object detection in large images using deep reinforcement learning." In The IEEE Winter Conference on Applications of Computer Vision, pp. 1824-1833. 2020. (IEEE WACV 2020)
Sarukkai Vishnu, Anirudh Jain, Burak Uzkent, and Stefano Ermon. "Cloud Removal from Satellite Images using Spatiotemporal Generator Networks." In The IEEE Winter Conference on Applications of Computer Vision, pp. 1796-1805. 2020. (IEEE WACV 2020)
M Rustowicz R, Cheong R, Wang L, Ermon S, Burke M, Lobell D. Semantic Segmentation of Crop Type in Africa: A Novel Dataset and Analysis of Deep Learning Methods. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops 2019 (pp. 75-82).
Oshri B, Hu A, Adelson P, Chen X, Dupas P, Weinstein J, Burke M, Lobell D, Ermon S. Infrastructure quality assessment in Africa using satellite imagery and deep learning. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining 2018 Jul 19 (pp. 616-625). ACM.
Wang AX, Tran C, Desai N, Lobell D, Ermon S. Deep transfer learning for crop yield prediction with remote sensing data. In Proceedings of the 1st ACM SIGCAS Conference on Computing and Sustainable Societies 2018 Jun 20 (p. 50). ACM.
Perez A, Yeh C, Azzari G, Burke M, Lobell D, Ermon S. Poverty prediction with public landsat 7 satellite imagery and machine learning. arXiv preprint arXiv:1711.03654. 2017 Nov 10.
Sheehan E, Uzkent B, Meng C, Tang Z, Burke M, Lobell D, Ermon S. Learning to Interpret Satellite Images Using Wikipedia. arXiv preprint arXiv:1809.10236. 2018 Sep 19.
Uzkent B, Sheehan E, Meng C, Tang Z, Burke M, Lobell D, Ermon S. Learning to Interpret Satellite Images Using Wikipedia in Global Scale. International Joint Conference on Artificial Intelligence, 2019.
Sheehan E, Meng C, Matthew Tan, Uzkent B, Neal J., Burke M, Lobell D, Ermon S. Predicting Economic Development using Geolocated Wikipedia Articles. 25th Acm Sigkdd Conference on Knowledge Discovery and Data Mining, 2019.
Hu W, Patel JH, Robert ZA, Novosad P, Asher S, Tang Z, Burke M, Lobell D, Ermon S. Mapping Missing Population in Rural India: A Deep Learning Approach with Satellite Imagery.
Jean N, Wang S, Azzari G, Lobell D, Ermon S. Tile2Vec: Unsupervised representation learning for remote sensing data. arXiv preprint arXiv:1805.02855. 2018 May 8.
Apollo Kaneko, Thomas Kennedy, Lantao Mei, Christina Sintek, Marshall Burke, Stefano Ermon, David Lobell, "Deep Learning for Crop Yield Prediction in Africa." International Conference on Machine Learning AI for Social Good Workshop 2019.