Institute for Data, Systems and Society (IDSS): Overarching Challenges

Cascading failures in power grids resulted in blackouts in New York City following Hurricane Sandy.

Resilience and Systemic Risk -- The complex interaction between physical systems, cyber layers, and humans has created new difficulties in achieving efficiency and optimality while maintaining reliability in contemporary, complex, man-made systems. “Systemic risk” is a term used to describe fragility in systems that can result in a cascade of failures in interconnected systems due to the aggregation of small disturbances or large disruptions. IDSS research will aim to create a foundational science that allows for measuring, predicting, and containing systemic risk. This theoretical development will emerge from an in-depth understanding of systemic risk in the application of complex systems that span a wide range of diversity in terms of interconnections, degree of automation and decentralization, time scales of both evolution and decisions, and the role of humans in their operations. The ability to adapt and reconfigure is the main attribute of a resilient system.

System Design and Architecture -- Good architecture is easy to recognize in retrospect, but elusive to predict or design. As we transform the nation’s power grid, move towards smart transportation systems, or enable global, real-time data exchange in financial markets, we must create the mathematical underpinning of network architecture and systematic methods to develop and evaluate design choices and algorithms. IDSS will incorporate foundational theory, practical algorithms, and concrete applications to guide us to a framework for secure and efficient architectures.

Social Welfare, Sustainability, and Policy -- Achieving social welfare, resilience, and adaptation across the domains of ecology, economics, politics, and culture requires systematic evaluations of public and scientific innovations. The key to addressing challenges is to enable both quantitative analysis and design of policy. Given the complexity of societal applications, as well as the long time frame needed to plan for resilience and sustainability, deriving appropriate models is critical to well-informed public policy.

Data-to-Decisions -- How do we effectively use theory, statistical models, and algorithms to make better decisions? Deriving actionable insights from data can be challenging , and there are important technical considerations as well as management and ethical issues which we must take into account. While the identification of models for decision making has been a hallmark of decision theory, the sheer magnitude, heterogeneity, and structure of modern data sets creates new challenges. In addition, we need to address issues such as privacy and security while still keeping the data informative and useful. In developing new analytical methods and approaches, we must understand the constraints, priorities, and context in which decisions will be made. We must address privacy issues when it comes to personal data, and the use of this data by different parties. We must learn to convey meaningful stories from data by using new tools for data exploration and visualization. We must also understand the inherent biases in the data and the limits of our models, in order to establish trust and enable others to make informed decisions, in particular when the outcomes of these decisions have a direct impact on the quality of people’s lives.