We will present a very general framework for unconstrained stochastic optimization which is based on standard trust region framework using  random models. In particular this framework retains the desirable features such step acceptance criterion, trust region adjustment and ability to utilize of second order models. We make assumptions on the stochasticity that are different from the typical assumptions of stochastic and simulation-based optimization. In particular we assume that our models and function values satisfy some good quality conditions with some probability fixed, but can be arbitrarily bad otherwise. We will analyze the convergence and convergence rates of this general framework and discuss the requirement on the models and function values. We will will contrast our results with existing results from stochastic approximation literature. We will finish with examples of applications arising the area of machine learning. 
 

Abstract: Triangulations of surfaces serve as important examples for cluster theory, with the natural operation of “diagonal flips” encoding mutation in cluster algebras and categories. In this talk we will focus on the combinatorics of mutation on marked surfaces with infinitely many marked points, which have gained importance recently with the rising interest in cluster algebras and categories of infinite rank. In this setting, it is no longer possible to reach any triangulation from any other triangulation in finitely many steps. We introduce the notion of mutation along infinite admissible sequences and show that this induces a preorder on the set of triangulations of a fixed infinitely marked surface. Finally, in the example of the completed infinity-gon we define transfinite mutations and show that any triangulation of the completed infinity-gon can be reached from any other of its triangulations via a transfinite mutation. The content of this talk is joint work with Karin Baur.

Categorification of knot polynomials -- Daniele Celoria

Classically, the most powerful and versatile knot invariants take the form of polynomials. These can usually be defined by simple recursive equations, known as skein relations; after giving the main examples of polynomial knot invariants (Alexander and Jones polynomials), we are going to informally introduce categorifications. Finally we are going to present the Knot Floer and the Khovanov homologies, and show that they provide a categorification of the aforementioned polynomial knot invariants.

Network science for online social media: an x-ray or a stethoscope for society -- Mariano Beguerisse

The abundance of data from social media outlets such as Twitter provides the opportunity to perform research at a societal level at a scale unforeseen. This has spurred the development of mathematical and computational methods such as network science, which uses the formalism and language of graph theory to study large systems of interacting agents. In this talk, I will provide a sketch of network science and its application to study online social media. A number of different networks can be constructed from Twitter data, which can be used to ask questions about users, ranging from the structural (an 'x-ray' to see how societies are connected online) to the topical ('stethoscope' to feel how users interact in the context of specific event). I will provide concrete examples from the UK riots of 2011, applications to medical anthropology, and political referenda, and will also highlight distinct challenges such as the directionality of connections, the size of the network, the use of temporal information and text, all of which are active areas of research.

Do you find yourself agreeing to things when actually you want more – or less? In this session we will look at how to be clear about what you want, and how to use assertiveness and negotiation skills and strategies to achieve win-win outcomes when working with others. 

This session will look at the range of courses available to early career researchers and graduate students from MPLS. It will also discuss the benefits of training and development for researchers and how it can help you in enhancing your career inside and outside academia.
 

Featuring
Peter Grindrod, Director of the Oxford-Emirates Data Science Lab, Oxford Mathematical Institute

Geraint Lloyd, Senior Software Engineer, Schlumberger

[[{"fid":"45910","view_mode":"media_square","fields":{"format":"media_square","field_file_image_alt_text[und][0][value]":"Geraint Lloyd","field_file_image_title_text[und][0][value]":"Geraint Lloyd"},"type":"media","attributes":{"alt":"Geraint Lloyd","title":"Geraint Lloyd","height":"258","width":"258","class":"media-element file-media-square"}}]]

Mick Pont, VP Research and Development, Numerical Algorithms Group (NAG)

[[{"fid":"45911","view_mode":"media_square","fields":{"format":"media_square","field_file_image_alt_text[und][0][value]":"Mick Pont","field_file_image_title_text[und][0][value]":"Mick Pont"},"type":"media","attributes":{"alt":"Mick Pont","title":"Mick Pont","height":"258","width":"258","class":"media-element file-media-square"}}]]

Anna Railton, Technical Staff, Smith Institute

[[{"fid":"45912","view_mode":"media_square","fields":{"format":"media_square","field_file_image_alt_text[und][0][value]":"Anna Railton","field_file_image_title_text[und][0][value]":"Anna Railton"},"type":"media","attributes":{"alt":"Anna Railton","title":"Anna Railton","height":"258","width":"258","class":"media-element file-media-square"}}]]
Michele Taroni, Senior Project Manager, Roxar

[[{"fid":"45913","view_mode":"media_square","fields":{"format":"media_square","field_file_image_alt_text[und][0][value]":"Michele Taroni","field_file_image_title_text[und][0][value]":"Michele Taroni"},"type":"media","attributes":{"alt":"Michele Taroni","title":"Michele Taroni","height":"258","width":"258","class":"media-element file-media-square"}}]]

A continuum of expanders -- David Hume

Expanders are a holy grail of networking; robustly connected networks of arbitrary size which require minimal resources. Like the grail, they are also notoriously difficult to construct. In this talk I will introduce expanders, give a brief overview of just a few aspects of their diverse history, and outline a very recent result of mine, which states that there are a continuum of expanders with fundamentally different large-scale geometry.

What makes cities successful? A complex systems approach to modelling urban economies -- Neave O'Clery

Urban centres draw a diverse range of people, attracted by opportunity, amenities, and the energy of crowds. Yet, while benefiting from density and proximity of people, cities also suffer from issues surrounding crime, congestion and density. Seeking to uncover the mechanisms behind the success of cities using novel tools from the mathematical and data sciences, this work uses network techniques to model the opportunity landscape of cities. Under the theory that cities move into new economic activities that share inputs with existing capabilities, path dependent industrial diversification can be described using a network of industries. Edges represent shared necessary capabilities, and are empirically estimated via flows of workers moving between industries. The position of a city in this network (i.e., the subnetwork of its current industries) will determine its future diversification potential. A city located in a central well-connected region has many options, but one with only few peripheral industries has limited opportunities.

We develop this framework to explain the large variation in labour formality rates across cities in the developing world, using data from Colombia. We show that, as cities become larger, they move into increasingly complex industries as firms combine complementary capabilities derived from a more diverse pool of workers. We further show that a level of agglomeration equivalent to between 45 and 75 minutes of commuting time maximizes the ability of cities to generate formal employment using the variety of skills available. Our results suggest that rather than discouraging the expansion of metropolitan areas, cities should invest in transportation to enable firms to take advantage of urban diversity.

This talk will be based on joint work with Eduardo Lora and Andres Gomez at Harvard University.