Date 
Speaker 
Title 
1/17/18 
Cheng Ouyang
University of Illinois at Chicago 
Local density estimate for a hypoelliptic SDE
Abstract
In a series of three papers in the 80's, Kusuoka and Stroock developed a probabilistic program in order to obtain sharp bounds for the density function of a hypoelliptic SDE driven by a Brownian motion. We aim to investigate how their method can be used to study rough SDEs driven by fractional Brownian motions. In this talk, I will outline Kusuoka and Stroock's approach and point out where the difficulties are in our current setting.
The talk is based on an ongoing project with Xi Geng and Samy Tindel.

1/24/18 
Aaron Yip
Purdue University 
Long Time Behavior of Brownian Motion in Tilted Periodic Potentials
Abstract
We will investigate various limits concerning the long time average
velocity ${\scriptsize V}$ of a Brownian particle diffusing on a periodic potential. The
prototype model is Langevin dynamics which incorporates inertia (mass) and
friction. The key feature of the current work is the consideration of an
additional macroscopic tilt, ${\scriptsize F}$. The goal is to understand how the average
velocity ${\scriptsize V}$ depends on ${\scriptsize F}$. Interesting thresholds for the value of ${\scriptsize F}$ can be
obtained, in particular under the limit of vanishing friction and noise.
Using the averaging theory of FreidlinWentzel, the current work provides
rigorous mathematical justification of some formulas obtained by Risken.
An earlier, fundamental result by Tanaka for Brownian particle diffusing
on a random (Brownian) potential will also be discussed.
This is a joint work with Liang Cheng.

1/31/18 
Donatella Danielli
Purdue University 
Obstacle problems for nonlocal operators
Abstract
In this talk we will discuss stationary and evolution obstacle problems defined by a class of nonlocal operators that are not stablelike
and may have supercritical drift. Our goal is to establish existence, uniqueness, and regularity of viscosity solutions. In particular, we will give sufficient conditions on the coefficients of the operator to obtain Hölder and Lipschitz continuous solutions. The class of nonlocal operators that we
consider include nonGaussian asset price models widely used in mathematical finance, such as
Variance Gamma Processes and Regular Lévy Processes of Exponential type.

2/7/18 
Xi Geng
Carnegie Mellon University 
Long time asymptotics of heat kernels and its application to Brownian winding numbers
Abstract
In this talk, we investigate the exact long time asymptotics of the heat kernel on abelian covering spaces over Riemannian manifolds with Dirichlet or Neumann boundary conditions. As an application, we establish a Gaussian type central limit theorem for the winding number of reflected Brownian motion in a planar domain with multiple holes. In particular, this is a refinement of a classical result by Toby and Werner, in which a law of large numbers was obtained. This is a recent joint work with Gautam Iyer.

2/14/18 
Rodrigo Bañuelos
Purdue University 
The $p$norm of the discrete Hilbert transform
Abstract
The discrete Hilbert transform was introduced by David Hilbert at the beginning of the 20th century. In 1925, Marcel Riesz proved the ${\scriptsize L^p}$ boundedness, ${\scriptsize p\in(1,\infty)}$, of the continuous version of this operator (also called the conjugate function). From this, he deduced the same result for the discrete version. In 1926, Titchmarsh gave a new proof of Riesz’s result and claimed that in addition the two operators have the same ${\scriptsize p}$norms. The following year, Titchmarsh pointed out that his argument for the equality of the norms was incorrect. The problem of equality has been a longstanding conjecture since. In this talk we describe a proof of this conjecture. This is joint work with Mateusz Kwasnicki of Wroclaw University.

2/21/18 
Takashi Owada
Purdue University Statistics Department 
Subtree counts on hyperbolic random geometric graphs
Abstract
We shall consider a geometric graph model on the "hyperbolic" space, which is characterized by a negative Gaussian curvature. Among several equivalent models representing the hyperbolic space, we treat the most commonly used ddimensional Poincare ball model. One of the main characteristics of geometric graphs on the hyperbolic space is treelike hierarchical structure. From this viewpoint, we discuss the asymptotic behavior of subtree counts. The spatial distribution of subtrees is crucially determined by an underlying curvature of the space. For example, if the space gets flatter and closer to the Euclidean space, subtrees are asymptotically scattered near the center of the Poincare ball. On the contrary, if the space becomes 'more hyperbolic' (i.e., more negatively curved), the distribution of trees is asymptotically determined by those concentrated near the boundary of the Poincare ball. We investigate the asymptotics of the expectation and variance of subtree counts. Moreover, we prove the corresponding central limit theorem as well.
This is joint work with Yogeshwaran D. at Indian Statistical Institute.

2/28/18 
Tom Sellke
Purdue University Statistics Department 
Theorems for the frontier of onedimensional branching diffusions
Abstract
This talk will discuss results in a joint paper of mine with Steve Lalley from 1992. Suppose a particle, starting at position 0, moves according to a diffusion process on the real line. Suppose also that this particle emits daughter particles according to a branching process whose instantaneous rate can depend on location, though not on time. The daughter particles move independently according to the same diffusion process, starting at their points of birth, and in turn emit their own daughters according to the same branching process. The simplest special case of this situation is standard branching Brownian motion, with the rate of reproduction not depending on location. Let ${\scriptsize R_t}$ be the position of the rightmost particle at time t, and let ${\scriptsize m_t}$ be the median of ${\scriptsize R_t}$. In 1937, Kolmogorov, Petrovskii, and Piskunov showed that, for standard branching Brownian motion, ${\scriptsize (m_t)/ t}$ converges to SQRT(2) and that ${\scriptsize (R_t  m_t)}$ converges in distribution to a nondegenerate limiting distribution.It turns out that results like those proved by Kolmogorov, et al, hold in great generality for onedimensional branching diffusions. If the branching diffusion is 'recurrent' (in the sense that the initial position is revisited at arbitrarily large times by some particle), and if space is rescaled so that m_t grows linearly, then ${\scriptsize (R_t  m_t)}$ converges in distribution to a locationmixture of extreme value distributions. We also have the Andy Warhol Theorem, according to which every particle ever born has a descendant in the lead at some point in the future.

3/7/18 
Zachary Selk
Purdue University 
Introduction to Rough Paths Theory
Abstract
Rough Paths Theory was developed in the 1990s by Terry Lyons. Recently, Martin Hairer used rough paths to create solution theories to stochastic PDEs such as KPZ equation and introduced a generalization known as regularity structures, for which he won the Field's medal in 2014. Rough paths allow us to define pathwise solutions to controlled differential equations driven by signals with low Hölder regularity, such as fractional Brownian motion. Classically, if the driving signal is nice enough, we may define a RiemannStieltjes integral. In this talk, we discuss the limits of classical Young/Stieltjes integration, define rough paths and use them to define a rough integral against a controlled rough path, which we take to be the solution of a controlled differential equation.

3/21/18 
Harsha Honnapa
Purdue University Industrial Engineering 
Analytical Approximations for Nonhomogeneous Markov Chains
Abstract
Performance analysis of engineered systems, such queues and inventories, typically use Markov processes as representative models. Most 'real world' systems display timevarying behavior; for example timeofday and dayofweek effects in traffic models of queues. On the other hand, the most widely used models in practice have strong stationarity and time homogeneity assumptions. This is with good reason: homogeneous stochastic models are far easier to use for computation and control/optimization purposes. It would, therefore, be very useful to have analytical approximations to nonhomogeneous performance measures that do not have any greater computational complexity than the homogeneous counterparts. In this talk I will present some new results on analytical approximations for 'slowly changing' Markov chains, as well as older results on uniform acceleration approximations to nonhomogeneous Markov chains. This is joint work with Peter W. Glynn and Zeyu Zheng at Stanford University.

3/28/18 
JuYi Yen
University of Cincinnati 
TBA
Abstract
Abstract coming soon.

4/4/18 
Vivian Healey
University of Chicago 
TBA
Abstract
Abstract coming soon.

4/11/18 
Jorge León
Cinvestav (Mexico) 
Malliavin calculus for Levy processes
Abstract
The purpose of this talk is to give an idea of the
Malliavin calculus techniques for Lévy processes, based on the socalled annihilation
operator and its adjoint. We will see that this important
stochastic analisis tool allows us to consider problems that cannot be studied
by means of the classical Itô calculus, although me might be in an adapted context with respect to the underlying filtration.

4/18/18 
Tom Sellke
Purdue University Statistics Department 
Theorems for the frontier of onedimensional branching diffusions (2)
Abstract
This talk will discuss results in a joint paper of mine with Steve Lalley from 1992. Suppose a particle, starting at position 0, moves according to a diffusion process on the real line. Suppose also that this particle emits daughter particles according to a branching process whose instantaneous rate can depend on location, though not on time. The daughter particles move independently according to the same diffusion process, starting at their points of birth, and in turn emit their own daughters according to the same branching process. The simplest special case of this situation is standard branching Brownian motion, with the rate of reproduction not depending on location. Let ${\scriptsize R_t}$ be the position of the rightmost particle at time t, and let ${\scriptsize m_t}$ be the median of ${\scriptsize R_t}$. In 1937, Kolmogorov, Petrovskii, and Piskunov showed that, for standard branching Brownian motion, ${\scriptsize (m_t)/ t}$ converges to SQRT(2) and that ${\scriptsize (R_t  m_t)}$ converges in distribution to a nondegenerate limiting distribution.It turns out that results like those proved by Kolmogorov, et al, hold in great generality for onedimensional branching diffusions. If the branching diffusion is 'recurrent' (in the sense that the initial position is revisited at arbitrarily large times by some particle), and if space is rescaled so that m_t grows linearly, then ${\scriptsize (R_t  m_t)}$ converges in distribution to a locationmixture of extreme value distributions. We also have the Andy Warhol Theorem, according to which every particle ever born has a descendant in the lead at some point in the future.

4/25/18 
Panki Kim
Seoul Nat. University (Korea) 
Heat kernel estimates for symmetric jump processes with general mixed polynomial growths
Abstract
In this talk, we discuss transition densities of pure jump symmetric Markov processes in ${\scriptsize R^d}$, whose jumping kernels are comparable to radially symmetric functions with general mixed polynomial growths. Under some mild assumptions on their scale functions, we establish sharp twosided estimates of transition densities (heat kernel estimates) for such processes.
This is a joint work with Joohak Bae, Jaehoon Kang and Jaehun Lee.
