Medical imaging is a key diagnostic tool, and is paramount for disease detection and for patient monitoring during ongoing care. Often, to reduce the amount of radiation that a patient is subjected to, there is a strong incentive to consider image reconstruction from incomplete sets of measurements, and so the imaging process is formulated as a compressed sensing problem.

In this talk, we will focus on compressed sensing for digital tomosynthesis (DTS), in which three-dimensional images are reconstructed from a set of two-dimensional X-ray projections. We first discuss a reconstruction approach for static bodies, with a particular interest in the choice of basis for the image representation. We will then focus on the need for accurate image reconstructions when the body of interest is not stationary, but is undergoing simple motion, discussing two different approaches for tackling this dynamic problem.

I will describe a novel algorithm for computing the Walsh Hadamard Transform (WHT) which consists entirely of Haar wavelet transforms. The algorithm shares precisely the same serial complexity as the popular divide-and-conquer algorithm for the WHT. There is also a natural way to parallelize the algorithm which appears to have a number of attractive features.

In laboratories around the world, scientists use magnetic stirrers to mix solutions and dissolve powders. It is well known that at high drive rates the stir bar jumps around erratically with poor mixing, leading to its nick-name 'flea'. Investigating this behaviour, we discovered a state in which the flea levitates stably above the base of the vessel, supported by magnetic repulsion between flea and drive magnet. The vertical motion is oscillatory and the angular motion a superposition of rotation and oscillation. By solving the coupled vertical and angular equations of motion, we characterised the flea’s behaviour in terms of two dimensionless quantities: (i) the normalized drive speed and (ii) the ratio of magnetic to viscous forces. However, Earnshaw’s theorem states that levitation via any arrangement of static magnets is only possible with additional stabilising forces. In our system, we find that these forces arise from the flea’s oscillations which pump fluid radially outwards, and are only present for a narrow range of Reynold's numbers. At slower, creeping flow speeds, only viscous forces are present, whereas at higher speeds, the flow reverses direction and the flea is no longer stable. We also use both the levitating and non-levitating states to measure rheological properties of the system.

A well known result by Schaeffer (1973) shows that generic solutions to a scalar conservation law are piecewise smooth, containing a finite family of shock curves.

In this direction, it is of interest to find other classes of nonlinear hyperbolic equations where nearly all solutions (in a Baire category sense) are piecewise smooth, and classify their singularities.

The talk will mainly focus on conservative solutions to the nonlinear variational wave equation $u_{tt} - c(u)(c(u) u_x)_x = 0$. For an open dense set of $C^3$ initial data, it is proved that the conservative solution is piecewise smooth in the $t - x$ plane, while the gradient $u_x$ can blow up along  finitely  many characteristic curves. The analysis relies on a variable transformation which reduces the equation to a semilinear system with smooth coefficients, followed by an application of Thom's transversality theorem.   

A detailed description of the solution profile can be given, in a neighborhood of every singular point and every singular curve.

Some results on structurally stable singularities have been obtained  also for dissipative solutions, of the above wave equation. Recent progress on the Burgers-Hilbert equation, and related open problems, will also be discussed.

These results are in collaboration with Geng Chen, Tao Huang, Fang Yu, and Tianyou Zhang.

Current technological progress has raised concerns about automation of tasks performed by workers resulting in job losses. Previous studies have used machine learning techniques to compute the automation probability of occupations and thus, studied the impact of automation on employment. However, such studies do not consider second-order effects, for example, an occupation with low automation probability can have a  surplus of labor supply due to similar occupations being automated. In this work, we study such second-order effects of automation using a network approach.  In our network – the Job Space – occupations are nodes and edges link occupations which share a significant amount of work activities. By mapping employment, automation probabilities into the network, and considering the movement of workers, we show that an occupation’s position in the network may be crucial to determining its employment future.