Lily Xu
Columbia University
Selective interventions for information acquisition: Incentives and audits for biodiversity
Lily Xu · Columbia University
The biodiversity crisis threatens one million species with extinction and jeopardizes the ecosystems on which over half of global GDP depends. Increasingly, efforts to conserve and restore biodiversity are made through distributed efforts, from contractor programs to remove invasive species to smallholder landowners restoring habitats. Because these activities are geographically dispersed and difficult to observe directly, obtaining reliable information about what is happening on the ground can be costly. This raises a central question: where, when, and how should we intervene to obtain the information that matters most? With examples from managing invasive Burmese pythons in Florida and monitoring for biodiversity credits in Kenya, we’ll discuss how active learning, sequential inference, and incentive design can help focus limited resources on the information most consequential for downstream decisions.
About the speaker
Lily Xu holds the Sun-Wu Assistant Professorship in Columbia’s Industrial Engineering and Operations Research department and is affiliated with Computer Science. She develops AI tools for environmental and planetary health problems, especially biodiversity conservation, drawing on machine learning, optimization, and causal inference. She co-directs the EAAMO research initiative and works with conservation organizations as the SMART Partnership’s AI Lead and an INVERSA scientific advisor. She earned her computer science doctorate at Harvard and subsequently held a postdoctoral position at Oxford’s Leverhulme Centre for Nature Recovery.