Research
We develop AI methods that create real-world impact for people and the planet.
Our vision
Planetary health is an emerging field which recognizes that human wellbeing and the health of our planet are inextricably linked. Addressing today's global polycrisis—from biodiversity loss and climate change to food insecurity and infectious disease—demands urgent, time-sensitive action. Unfortunately, the set of possible interventions is vast and available data remains incomplete, making it difficult to anticipate the long-term impacts of our decisions.
The caPP Lab seeks to bridge this gap by developing data-driven methods that are both efficient enough to meet the scale of our planet's most pressing challenges and robust to the uncertainty intrinsic to these high-stakes settings. We uncover research questions through close engagement with decision makers on the ground, identify the barriers they face in translating data to action, then design novel AI algorithms to turn messy data into reliable decisions.
We envision a future in which fit-for-purpose AI systems are intentionally designed to augment, not replace, the expertise of practitioners. By advancing the state-of-the-art in sequential decision-making and working in close partnership with NGOs, industry, and intergovernmental organizations, we strive to ensure that every algorithmic advance translates into a more sustainable future for both people and the planet.
Topics
Our research sits at the intersection of machine learning, optimization, and causal inference, developing new methodology inspired by practical problems in planetary health.
- Sequential decision making: multi-armed bandits, reinforcement learning, robust planning, game theory
- Machine learning: ML + causality; ML + optimization; data science; learning in sparse, noisy settings
- AI for social impact: AI for conservation, bridging research and practice, community building
- Biodiversity conservation: nature financing, protected area management, ranger-based monitoring
Invasive species management
Coordinating agents to effectively respond to invasive animals, with a focus on the greater Florida Everglades.
Disaster relief
Equitable and robust resource allocation for sudden-onset disasters, with a focus on earthquake response in Turkey.
Optimal monitoring for biodiversity credits
Efficient data collection for ecological monitoring to increase the scale and viability of biodiversity credit programs.
Illegal fishing
Planning strategic enforcement to combat illegal, unreported, and unregulated (IUU) fishing in Chile.
Maternal health
Sequential resource allocation in low-resourced public health settings where limited interventions must be allocated across many beneficiaries over time.
Wildlife conservation
Predictive models and patrol planning algorithms (across online learning and robust reinforcement learning) for anti-poaching to enable more effective protected area management.
Partnerships
We bridge research and practice through collaboration with NGOs, governments, and intergovernmental organizations.
- 🐘SMART Partnership, now known as SERCA
Building computational solutions for effective conservation management across protected areas worldwide. - 🐍INVERSA
Harnessing market forces to protect threatened species by creating value from invasive animals. - 🐦Natural State
Pioneering novel financial mechanisms, capacity building, and impact monitoring to catalyze funding for large-scale landscape restoration and conservation. - 🤝Hayata Destek
Helping disaster-affected communities meet their basic needs and rights, based in Istanbul. - 🤱ARMMAN
Leveraging mobile health (mHealth) technology to create cost-effective, scalable, gender transformative, non-linear, systemic solutions to reduce maternal and child mortality and morbidity in India.