The Complex Systems Science (CSS) Laboratory is an interdisciplinary lab in the Whitaker Biomedical Engineering Institute at the Johns Hopkins University. The main objective of the work performed in this lab is to facilitate research in the general areas of modeling, identification, and analysis of complex networked systems. Key research activities are focused on developing state-of-the-art methodologies for model-based identification and analysis of complex nonlinear interaction networks from incomplete and noisy observations. Researchers in the lab employ rigorous theoretical and computational techniques for estimating the structural and dynamic properties of complex interaction networks by state-of-the-art identification and model selection methodologies, and for studying network robustness via probabilistic sensitivity analysis techniques. Although the main applications of interest are in problems of computational systems biology, applications in other areas (such as epidemiology) are also considered.

What are complex networked systems?

Complex networks, consisting of interacting elements linked together with processing units, are ubiquitous across many disciplines in modern science and engineering. Examples are listed in the following table:

Network	Components ( Interacting Elements)	Interactions (Processing Units)
biochemical	molecules	chemical reactions
biological neural	neurons	synapses
epidemiological	healthy, infected individuals	infectious contacts
world wide web	web pages	hyper links
trophic	predators, prey	consumption
power grid	electrical generators	power lines, substation transformers
collaborative	people	collaborations
social	people	contacts
internet	computers	routers, ethernet cables

Complex network science involves the study of such networks using advanced mathematical and statistical techniques as well as sophisticated computational approaches. The components and interactions in these networks need not be complex individually, but the large size and complicated structure, or topology, of such networks is what makes these systems to be "complex."

Why study complex networked systems?

Despite the enormous diversity of complex networked systems, researchers have been amazed by the striking commonalities between them. For decades, specific examples of complex networks have been studied intensely within their unique disciplines. Recently, however, there has been a surge of interest in studying complex networked systems in their own right by bridging the gap between diverse disciplines and by developing unifying mathematical, statistical, and computational approaches for their description and analysis.

Many interesting and challenging questions of modern science relate to complex networked systems:

• How does human cognition arise from the firing of interconnected neurons in the brain?

• How can we prevent a disease outbreak from spreading to a pandemic, or minimize the damage of a biological attack?

• How do the genome and proteome interact to create homeostatic and epigenetic behavior in cells? How can we repair or prevent diseased states at the molecular level?

• What is the best search algorithm for the WWW? Can we design a better physical internet or protocol than IPv6?

• What can be done to maximize the robustness of our public infrastructure to random or targeted failures and attacks?

What is our approach?

We are developing a general methodology for model-based identification and analysis of complex nonlinear interaction networks from incomplete and noisy observations. We are employing rigorous theoretical and computational techniques for estimating the structural and dynamic properties of complex interaction networks by state-of-the-art identification and model selection methodologies, and for studying network robustness via probabilistic sensitivity analysis. We use fundamental laws of physics (such as thermodynamics) to appropriately constrain the systems under consideration and focus our identification and analysis methodologies only on physically realizable networks. Sensitivity analysis is used from start to finish: in model reduction to ease inference, in experimental design to enhance scientific discovery, and in final system analysis for controlled intervention.