## Events Search and Views Navigation

## March 2021

### Eindhoven SPOR Seminar

Christian Brownlees (UPF) Community Detection in Partial Correlation Network Models We introduce a class of partial correlation network models with a community structure for large panels of time series. In the model, the series are partitioned into latent groups such that correlation is higher within groups than between them. We then propose an algorithm that allows one to detect the communities using the eigenvectors of the sample covariance matrix. We study the properties of the procedure and establish its consistency.…

Find out more »## April 2021

### Eindhoven SPOR Seminar

Christopher Hojny (TU/e) Possibilities and Limitations of Symmetry Handling Cutting Planes Branch-and-bound is a well-established tool for solving combinatorial optimization problems. If the combinatorial problem contains symmetric structures (such as symmetric graphs or identical objects), however, branch-and-bound will also explore many symmetric, and thus unnecessary, subproblems during the solving process. To accelerate the solution procedure, a standard technique is to detect symmetries of the problem and to add inequalities (cutting planes) to the problem formulation that prevent the solver from…

Find out more »## May 2021

### Eindhoven SPOR Seminar

Matthieu Jonckheere (UBA) Distance learning using Euclidean percolation: Following Fermat's principle In unsupervised statistical learning tasks such as clustering, recommendation, or dimension reduction, a notion of distance or similarity between points is crucial but usually not directly available as an input. We proposed a new density-based estimator for weighted geodesic distances that takes into account the underlying density of the data, and that is suitable for nonuniform data lying on a manifold of lower dimension than the ambient space. The…

Find out more »### Eindhoven Stochastic Seminar

Shaojie Tang (UTD) Fast Adaptive Submodular Maximization In this paper, we study the non-monotone adaptive submodular maximization problem subject to a cardinality constraint. We first revisit the adaptive random greedy algorithm proposed in \citep{gotovos2015non}, where they show that this algorithm achieves a $1/e$ approximation ratio if the objective function is adaptive submodular and pointwise submodular. It is not clear whether the same guarantee holds under adaptive submodularity (without resorting to pointwise submodularity) or not. Our first contribution is to show…

Find out more »## June 2021

### Eindhoven Stochastic Seminar

Ahmad Abdi (LSE) Ideal matrices and dyadic linear programming A 0,1 matrix M is ideal if the set cover inequalities Mx>=1, together with nonnegativity constraints x>=0, define a polyhedron where every vertex is integral. The study of ideal matrices was initiated in the 1960s by Alfred Lehman and Ray Fulkerson, and continues to this day due to their intimate connections to Integer Programming, Combinatorial Optimization, and Graph Theory. Despite having being studied for about 60 years, several basic questions about…

Find out more »### YEQT XIV: "Load balancing and scheduling for distributed service systems"

Summary Distributed service systems are ubiquitous, from server farms for cloud computing and charging stations for electric vehicles, to checkout lines at supermarkets and ICU beds in hospitals. The size of many of these systems has exploded over the past few years, which has created new challenges in the design of control policies. These new challenges, such as scalability, data locality, and server and task heterogeneity, have fueled a renewed interest in the study of these systems, and accelerated the…

Find out more »### Eindhoven Stochastic Seminar

Alessandra Cipriani (TUDelft) The discrete membrane model on trees The discrete membrane model (MM) is a random interface model for separating surfaces that tend to preserve curvature. It is a close relative of the discrete Gaussian free field (DGFF), for which instead the most likely interfaces are those preserving the mean height. However working with the two models presents some key differences. In particular, a lot of tools (electrical networks, random walk representation of the covariance) are available for the…

Find out more »### Eindhoven SPOR Seminar

Tim van Erven (UvA) Highlights of Online Machine Learning Online machine learning algorithms process data sequentially, either because the data are inherently sequential or because the whole data set is too large to load into memory all at once (e.g. when training neural networks). I will introduce the formal setting and present several highlights of classical and recent progress in this area.

Find out more »## August 2021

### Workshop YEP XVII: "Interacting Particle Systems"

Format The workshop will take place in the week Aug 30 - Sep 3, 2021. The event will be online. Summary The theory of Interacting Particle Systems focuses on the dynamics of systems consisting of a large or infinite number of entities, in which the mechanism of evolution is random and follows simple, local rules. The topic had its beginnings in the 70s and 80s, motivated by Statistical Physics and fundamental problems from Probability Theory. It has since developed into…

Find out more »## September 2021

### Eindhoven SPOR Seminar

Matteo D’Achille (Université Paris-Est Créteil) One dimensional Euclidean Random Assignment Problems: anomalous scaling and critical hyperbolae A Euclidean Random Assignment Problem (ERAP in short) is defined as follows. Take two $n$-samples $\mathcal{B} =(b_i)_{i=1}^n$ (blue points) and $\mathcal{R}=(r_j)_{j=1}^n$ (red points) of i.i.d. random variables valued on a metric space $(\Omega,D)$ of dimension $d$ according to a prob. measure $\vu$ (called disorder). For a permutation $\pi : \mathcal{B} \rightarrow \mathcal{R}$, let $\mathcal{H}(\pi)=\sum_{i=1}^n D(b_i,r_{\pi(i)})^p$ be a Hamiltonian depending on the energy-distance exponent $p\in…

Find out more »