Header image

Thinking Machines wins Best Paper Award at NeurIPS 2020 ML4D Workshop

December 15, 2020 blog-post artificial-intelligence machine-learning geospatial research-papers satellite-imagery development government non-profit ngo

We are honored to share that our team presented in two NeurIPS 2020 workshops over the weekend--the largest gathering of AI researchers in the world. We are especially excited to have received the award for Best Paper in the Machine Learning for Development (ML4D) workshop!

Mapping New Informal Settlements with Artificial Intelligence

Since 2014, nearly 2 million Venezuelans have fled to Colombia to escape one of the largest humanitarian crises in modern history. With many of these displaced populations living in informal settlement areas across the country, locating migrant settlements over large territories to provide humanitarian aid can be a major challenge.

In the Artificial Intelligence for Humanitarian Assistance and Disaster Response (AI + HADR) workshop last December 13, we presented a novel approach for rapidly and cost-effectively detecting new and emerging informal settlements that have emerged between 2015 and 2020 in Colombia using machine learning and publicly accessible satellite imagery. A big thank you to our partners at iMMAP Colombia and USAID for the collaborative work and support throughout the project.

Check out the case study here and read our paper, titled ‘Mapping New Informal Settlements using Machine Learning and Time Series Satellite Images: An Application in the Venezuelan Migration’ here.

Interpretable Poverty Mapping with Machine Learning

Access to accurate, granular, and up-to-date poverty data is essential for humanitarian organizations to identify vulnerable areas for poverty alleviation efforts. Although recent works have shown success in combining computer vision and satellite imagery for poverty estimation, the cost of acquiring high-resolution images coupled with black box models can be a barrier for development organizations.

For the Machine Learning for the Developing World (ML4D) workshop held last December 12, our team presented an interpretable and cost-efficient approach to poverty estimation using machine learning and readily accessible data sources such as social media data, publicly available satellite images, and volunteered geographic information. For this project, we’d like to thank our partners at the UN Development Program and the Zero Extreme Poverty PH 2030 movement for helping make this work possible.

We are honored to have been invited to this year’s ML4D, along with teams from 20 other countries working on AI solutions that aim to improve resilience in developing countries. Aside from being chosen as one of the four contributed talks at the workshop, we are thrilled to have been awarded as Best Paper for this year.

Check out our poster here and read our paper, titled ‘Interpretable Poverty Mapping using Social Media Data, Satellite Images, and Geospatial Information’ here.

Want to reach out to us? Contact us below!


Querying Safely in the (Google) Cloud

Programming in BigQuery may feel very light, but do you have a safety net for when the cloud goes down?

Thinking Machines’ Climate Action Commitment

Being data-driven means taking climate action. We’re trying to find our niche in the climate fight.

Using Satellite Imagery and Nighttime Lights to Understand the Impact of COVID-19 across Southeast Asian Cities

In this third blog in our series with ADB, we zoomed into three Southeast Asian cities throughout the pandemic to see how nighttime lights have changed alongside lockdown measures and other key data sources.