Joseph Aylett-Bullock, PhD

Data Scientist and Researcher, UN Department of Peace Operations
Research Advisor, RiskEcon Lab, Courant Institute, NYU
Research Fellow, Institute for Data Science, Durham University

(Note: Before April 2021 I published under Joseph Bullock)

My work focuses on understanding, analysing and simulating complex systems through the development of mathematical modelling and machine learning techniques at the intersection of physics and computational social sciences. Specifically, I focus on supporting responses to peace & security and humanitarian crises.

My work has been used in active peace & security and humanitarian situations in: Bangladesh, Brazil, DRC, Mali, Mozambique, Nepal, Pakistan, Somalia, Syria, Turkey, Ukraine and others.

I am currently a Data Scientist and Researcher with the UN Department of Peace Operations where I lead on the technical work of the Addressing Mis- and Disinformation Unit. Mis/dis and malinformation is highly present in societies globally, severly affecting individuals, communities and operations of organisations such as the UN. Mis/dis and malinformation often serves as a catalyst for crises to evolve. Understanding the spread of such information within communities - including the types of narratives, actors involved and their behavioural patterns, as well as the effects on these communities - can enable decision makers to better respond to potentially harmful narratives. Further, such information can help anticipate occurances of mis/dis and malinformation, serving as an early warning system and allowing for the pre-bunking of narratives before they evolve. Mathematical and computational modelling and analysis techniques can aid in understanding these complex socio-political systems. In this area, my work focuses on developing and applying new methods for assessing mis/dis and malinformation in the context of peacekeeping which operates in often highly dynamic and under researched settings - with ongoing conflicts and highly variable rates of internet penetration making for a complex information landscape.

This work falls under a broader research umbrella of applying mathematical and computer modelling and analysis techniques to anticipate and response to peace & security and humanitarian crises - such situations are highly complex, especially given the propensity to have mutliple simultaneous, and often interacting, crises. Complex systems methods can help us understand these crises better and enable more informed preparation and contingency planning, mitigation measures, rapid response efforts, and medium-to-long-term recovery and resilience building. Mathematical and computational modelling and analysis methods such as agent-based approaches, natural language processing and machine learning techniques can be applied at different stages of the process. My research interests lie in piecing apart this complexity, developing new methods and models addressing specific components, and combining these together into broader mechanisms for supporting decision making and policy initiatives. In addition, a key component of this work is understanding and communicating uncertainites - including in the data, models, and overall decision making process.

Prior to this position, I lead the Data Science and AI team for the New York office of United Nations Global Pulse, the innovation lab of the UN Secretary-General to harness big data and emerging technologies for sustainable development and humanitarian action. Here, my work included: applying machine learning to remote sensing tasks leading to the first flood map produced by the UN; modelling the spread of diseases in the world’s largest refugee settlement to inform public health programming; simulating the spread of migrants to anticipate humanitarian aid needs; and developing a pipeline to collect and analyse data from radio stations in remote areas for social listening to inform humanitarian and public health interventions.

Much of this work is deeply informed by my background in Physics. I optained my PhD from the Institute for Particle Physics Phenomenology at Durham University, the national centre for phenomenological research, and was part of the Centre for Doctoral Training in Data Intensive Science and Institute for Data Science there. My PhD research focused on investigating ways to apply and develop machine learning and mathematical modelling techniques for problems partiple physics, and also how these approches can be brought into the humanitarian domain to help address challenges in crisis response and disaster relief. As specific case studies, I worked on constructing high-dimensional functional approximationsto be used in large-scale particle collision simualtion and performing uncertainty analysis, as well as utilising many of the numerical modelling techniques developed in this field for agent-based epidemic modelling.

I also hold the position of Research Advisor at the NYU Risk Econ Lab, part of the Courant Institute for Mathematical Sciences, where I co-advise several Masters students.

Approaching these complex peace & security and humanitarian challenges requires a wide variety of expertise and knowledge - including, economic, social, behavioural, political, geographical, computational, etc.. I enjoy working in multi-disciplinary teams with a range of backgrounds and have learnt, and continue to learn, a great deal from my colleagues in different departments and faculties. Much of my work has been done in collaboration these colleagues which is essential to making both theoretical and tangible change.

Details about my publications, talks and other activities can be found on the Research page.

Contact info:

joseph.bullock [at] un.org