I study complex biological and information systems, the scaling properties of networks, and the general rules governing the acquisition of energy and information in complex adaptive systems. My focus is on the efficiency of growth and information exchange in biological and computational networks, and how the size and topology of networks determine emergent system behavior. I draw insights, tools, and approaches from different disciplines in an effort to find unifying principles in the natural world.

A full list of downloadable publications is available on the Publications page, or in my CV (pdf updated Sept. 2018). 

Research in the Moses Biological Computation lab is focused in three areas:

Modeling the Adaptive Immune System

Organisms, societies, and computers are all complex systems whose behavior emerges from the interactions of components. For example, effective immune response requires coordinated interactions between billions of immune cells with no central point of control, just like the behavior of ant colonies emerges from distributed communication between millions of ants. My research, in collaboration with Dr. Judy Cannon in UNM Pathology, examines how interactions among cells, signals and structures of the immune system lead to effective search in complex environments. 

Recent papers model T cell search: how T cells move in the lymph node in the search for dendritic cells (Fricke 2016 plos comp bio); how T cell search is guided by chemical gradients and structures in the influenza-infected lung (Levin 2016, Mrass 2017), and a new method for quantifying spatial associations among cell types called Normalized Mutual Information (Tasmin et al 2018).

This collaborative work has been supported by the NSF Advancing Theory in Biology Program, the UNM Center for Evolutionary and Theoretical Immunology (CETI), the UNM Spatiotemporal Modeling Center, the DARPA CRASH program and subsequent SBIR, and the James S. McDonnell Foundation Complexity Scholars program.

Bio-inspired Swarm Robotics 

We have built swarms of Ant-like robots to mimic foraging behaviors of ants to search for and collect resources in ways that efficient, robust and scalable. Under research funded by the NASA Swarmathon, we have built 100 Swarmie robots, demonstrating that swarm robots can be moved from simulation into the real world. (Hecker 2015, Fricke et al 2016, Lu et al 2018 (1), Lu et al 2016 (2), Ackerman et al 2018). 

Scaling in Biology, Immunology, and Computation

In biology, bigger usually means slower. Mice live fast and die young, compared to elephants, with shorter lifespans, faster heartbeats, and faster reproduction. The Metabolic Theory of Ecology (MTE) explains that the systematic slowing of biological rates with body size results from the properties of resource distribution networks, such as the cardiovascular networks in vertebrates. Larger networks that are constrained by the volume of the organism they supply, are physically constrained to slow the delivery rate of metabolic materials, and this affects the physiology, life history and ecology of living systems. Several of my papers propose models to explain how resource distribution networks determine scaling (Banavar et al 2010, Moses 2016) while others reveal resulting life history patterns in growth and reproduction (Moses 2008, Hou 2008, Charnov et al 2007).

The one apparent biological exception to ‘bigger is slower’ is the immune system. I study how the distributed architecture and distributed computation of the immune system allows nearly scale-invariant search. I seek to understand the biological mechanisms of scalable distributed search both for basic understanding of immune function, and the consequent spread of disease, and as a framework for achieving scalable distributed search in human engineered systems (Banerjee 2017, Moses & Banerjee 2011, Banerjee et al 2010 (1) & Banerjee et al 2010 (2)).

I have also modified metabolic scaling theory to explain that information networks scale up differently than energy distribution networks. We are using this understanding to build scalable robot swarms (Moses et al. 2016, Moses et al 2008,  Lu et al 2018).