Applied Information Theory
Many physical systems undergoing critical transitions are not amenable to the standard methods of the theory of phase transitions. An example of this is the glass transition. We introduced a new dynamic lengthscale based on the mutual information, that scales unlike the commonly used dynamic lengthscale (Nature Communications, 2015). Our work provided important support for thermodynamically based theories over those which posit that the glass transition is driven by dynamics. This work was reported on several science news websites (altmetric score >200). Four years after the paper was published, there is increasing support for the thermodynamic — rather than dynamic — interpretation of the glass transition, and our paper may be seen as an important piece of evidence in support of this improved understanding.
In another example, I have studied the differentiation of stem cells (Royal Society Interface, 2018). The exact point of this transition from stem to differentiated cell is of great interest since it is the point of no return. Using information theory we could identify the point of transition for haematopoietic stem cells using single-cell genomic data. In a third example, we developed a new technique to reconstruct hidden Markov models from continuous-time data and identified the transition between conformational states of DNA junctions (PLoS One, 2012).
In current work we are developing information theoretic tools for the study of the stability of networks.
In another example, I have studied the differentiation of stem cells (Royal Society Interface, 2018). The exact point of this transition from stem to differentiated cell is of great interest since it is the point of no return. Using information theory we could identify the point of transition for haematopoietic stem cells using single-cell genomic data. In a third example, we developed a new technique to reconstruct hidden Markov models from continuous-time data and identified the transition between conformational states of DNA junctions (PLoS One, 2012).
In current work we are developing information theoretic tools for the study of the stability of networks.
