Research

I am a postdoctoral research scientist in the Complex Networks group at Aalto University. Previously, I was a graduate student in the David Reichman group at Columbia University. I have been here since August 2007. Previous to that, I was in the group of Peter Rossky at the University of Texas at Austin.

Projects

2014--present: Software development for scientists

During a group brainstorming session at Aalto, we decided that it would be smart to improve our efficiency at programming. Thus begun a series of software development tutorials for scientists. As time goes on, this is slowly turning into a full graduate-level course at Aalto.

2011--present: Network science and community detection

Through my glass work, I became interested in the field of community detection. In this field, one has large data of networked interactions: instead of the actual data being of primary interest, the pattern and forms of interactions are most interesting. The idea of community detection is to look within this data and find the logical components, presumably relatively strongly connected subsets.

However, this is a harder problem than it looks since this structure is often weak, and not very well defined. I found that, while these "community detection algorithms" should be good, in practice they did not give good results on many real networks. I began critically analyzing these different methods to understand (more mathematically than empirically) the network invariants they should detect. This allows me to better understand the range of usefulness of each method.

I have since expanded into more sociophysics topics. While my interest is the math behind the methods, I apply them to various large datasets, and am in essence a data scientist.

2008--2013: Lattice glass models

My research is focused on a theoretical understanding of the nature of glassy dynamics.

Despite scientists having studied the glass transition for decades, the theoretical underpinnings of it are not yet well understood. One dominant hypothesis is that of kinetically constrained models, that is, that glassy dynamics (huge slowdown in dynamics, non-exponential relaxation, and dynamic heterogeneity) are fundamentally caused by the blocking of transitions. Another, more recent and less well studied hypothesis is that of thermodynamically constrained models, that postulate that the blocking of configurations is the root cause of glassy dynamics.

Despite these two hypotheses existing, little has been done to distinguish between them. My work is focused on studying differences between these points of view. I have developed a new thermodynamic model, am comparing it and other thermodynamic models to many examples of kinetic models.

This should provide a basis for understand the benefits of the opposing viewpoints of the glass transition.

2007: Jagla solution models

Water is an unusual solvent in that small hydrophobic solutes become increasingly soluble as temperature decreases. Not many simple models can capture this property.

In order to attempt to understand this, I examined a very simple model of water, the Jagla model, which is essentially a hard sphere model, but with a "shoulder", a small repulsive region at short range. This adds behavior similar to the expansion of water upon freezing, and thus has the potential to explain the increased solubility at lower temperatures.

The standard explanation for increasing solubility (of small hydrophobic solutes) in water at decreasing temperature relates to the increasingly tetrahedral structure of water. Thus, if we have a spherically symmetric model which can reproduce the solubility behavior, we know that the behavior is not exclusively dependent on the tetrahedral structure of water.

This project was begun at a very late stage of my time at UT-Austin, and continued by my coworker.

2005--2007: Solvation of small solutes in water

I studied the potential of mean force of hydrogen and hydrophobic bonds in water, in an attempt to study the temperature dependence of both of these. The degree of statistical noise was significant compared to the temperature-induced variation, and my efforts to reduce this to an acceptable level were not successful.

2004--2005: Cold denaturation of proteins

I worked with a labmate studying the dynamics of proteins as they cold denatured.