BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Centre for Mathematical Sciences - ECPv4.5.4//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Centre for Mathematical Sciences
X-ORIGINAL-URL:http://math-sciences.org
X-WR-CALDESC:Events for Centre for Mathematical Sciences
BEGIN:VEVENT
DTSTART;TZID=Europe/London:20170531T130000
DTEND;TZID=Europe/London:20170531T140000
DTSTAMP:20170624T051651
CREATED:20170530T155702Z
LAST-MODIFIED:20170530T155702Z
UID:2490-1496235600-1496239200@math-sciences.org
SUMMARY:José Luis Bravo Trinidad (Extremadura)
DESCRIPTION:Some experiences of teaching with Sage\nJosé will give an informal description of how Sage has been used at the University of Extremadura to deliver courses. Sage is a free computer algebra program which is closely based on python with an increasingly wide user base amongst research mathematicians.
URL:http://math-sciences.org/event/jose-luis-bravo-trinidad-extremadura
LOCATION:2-5\, Kirkby Place\, Plymouth\, PL4 6DT\, United Kingdom
CATEGORIES:Pure Mathematics,Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=UTC+0:20170531T140000
DTEND;TZID=UTC+0:20170531T150000
DTSTAMP:20170624T051651
CREATED:20170419T141616Z
LAST-MODIFIED:20170601T183320Z
UID:2459-1496239200-1496242800@math-sciences.org
SUMMARY:Peter Richtarik (Edinburgh)
DESCRIPTION:Joint Theoretical Physics and Applied Mathematics Seminar\n\nTitle: Stochastic reformulations of linear systems and efficient randomized algorithms\n\nAbstract:\n\nWe propose a new paradigm for solving linear systems with a very large number of equations. In our paradigm\, the system is first reformulated into a stochastic problem\, and then solved with a suitable (typically randomized) algorithm. Our stochastic reformulation is flexible as it depends on a user-defined parameter in the form of a distribution defining an ensemble of random matrices. The choice of the distribution directly influences the “condition number” of the reformulation\, which leads to the novel concept of “randomized reconditioning”. We give necessary and sufficient conditions for the reformulation to be exact\, i.e.\, for the solution set of the stochastic problem to be identical to the solution set of the linear system. We also show that the reformulation can be equivalently seen as a stochastic optimization problem\, stochastically preconditioned linear system\, stochastic fixed-point problem and as a probabilistic intersection problem. For instance\, the condition number of the reformulation is equal to the condition number of the stochastically preconditioned linear system\, and to the condition number of associated with the Hessian of the objective function appearing the stochastic optimization reformulation. Further\, we propose and analyze basic\, parallel and accelerated stochastic algorithms for solving the reformulated problem\, with linear convergence rates. The methods have natural and sometimes surprising interpretations from the viewpoint of each of the four reformulations. For instance\, the methods can be interpreted as basic\, parallel and accelerated variants of stochastic gradient descent\, stochastic Newton descent\, stochastic projection method and stochastic fixed-point method. The complexity of the basic variants scales linearly with the condition number of the reformulation\, while the accelerated variants scale with the square root of the condition number. Moreover\, all our methods lend themselves to a natural dual interpretation as “stochastic subspace ascent” methods\, a novel class of optimization algorithms not analyzed before. Stochastic dual coordinate ascent and stochastic dual Newton ascent arise in special cases. We prove global linear convergence of all our algorithms. Further\, we highlight a close connection to recent algorithmic developments in machine learning through casting the problem as an instance of the Empirical Risk Minimization problem in a new regime not studied before.\n\nThe above development can be extended to matrix inversion. In particular\, we develop and analyze a broad family of stochastic/randomized algorithms for inverting a matrix\, with specialized variants maintaining symmetry and/or positive definiteness of the iterates. All methods in the family converge globally and linearly\, with explicit rates. In special cases\, we obtain stochastic block variants of several quasi-Newton updates\, including bad Broyden (BB)\, good Broyden (GB)\, Powell-symmetric-Broyden (PSB)\, Davidon-Fletcher-Powell (DFP) and Broyden-Fletcher-Goldfarb-Shanno (BFGS). Ours are the first stochastic versions of these updates shown to converge to an inverse of a fixed matrix. Through a dual viewpoint we uncover a fundamental link between quasi-Newton updates and approximate inverse preconditioning. Further\, we develop an adaptive variant of randomized block BFGS\, where we modify the distribution underlying the stochasticity of the method throughout the iterative process to achieve faster convergence. Further\, for rectangular and non-invertible matrices\, variants of our methods can be shown to converge to the Moore-Penrose pseudoinverse.\n\n \n\n
URL:http://math-sciences.org/event/peter-richtarik-edinburgh
CATEGORIES:Applied Mathematics,Seminars,Theoretical Physics
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=UTC+0:20170614T160000
DTEND;TZID=UTC+0:20170614T170000
DTSTAMP:20170624T051651
CREATED:20170227T095521Z
LAST-MODIFIED:20170510T075749Z
UID:2385-1497456000-1497459600@math-sciences.org
SUMMARY:Tom Philbin (Exeter)
DESCRIPTION:Controlling and understanding reflection and transmission\nThe propagation of electromagnetic waves in inhomogeneous materials remains a subject of intense interest and activity. This is partly because the challenge is not merely computational; there is also a need for improved analytical techniques. The issue of reflection by inhomogeneous materials illustrates the point well. It would be naively expected that inhomogeneity always causes some reflection and that this becomes very significant if the refractive index changes appreciably over a wavelength. But there are many inhomogeneous index profiles that have strictly zero reflection even when the geometrical-optics approximation is arbitrarily bad. I will describe techniques that explain some aspects of this interesting physics in a very visual manner. These techniques can be used to derive recipes for materials with interesting reflection and transmission properties. I also consider to what extent reflection can in principle be eliminated at all frequencies\, given fundamental physical constraints. This last question has direct application to the Casimir effect.
URL:http://math-sciences.org/event/tom-philbin-exeter
LOCATION:2-5 Kirkby Place\, Plymouth\, PL4 6DT\, United Kingdom
CATEGORIES:Seminars,Theoretical Physics
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=UTC+0:20170619T103000
DTEND;TZID=UTC+0:20170619T130000
DTSTAMP:20170624T051651
CREATED:20170530T160325Z
LAST-MODIFIED:20170530T160325Z
UID:2492-1497868200-1497877200@math-sciences.org
SUMMARY:Peter Newstead (Liverpool)
DESCRIPTION:Brill-Noether theory on nodal curves\nAbstract to follow.\n\nNote that this is two-hour seminar\, with a coffee break at 11.30.
URL:http://math-sciences.org/event/peter-newstead-liverpool
LOCATION:2-5\, Kirkby Place\, Plymouth\, PL4 6DT\, United Kingdom
CATEGORIES:Pure Mathematics,Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=UTC+0:20170623T133500
DTEND;TZID=UTC+0:20170623T143500
DTSTAMP:20170624T051651
CREATED:20170616T124658Z
LAST-MODIFIED:20170616T124731Z
UID:2508-1498224900-1498228500@math-sciences.org
SUMMARY:Jeff Ralph (RSS William Guy Lecturer for 2017-18)
DESCRIPTION:Joint RSS South West Local Group Meeting and Mathematics Year 9 Event:\n\nSpeaker: Jeff Ralph (RSS William Guy Lecturer for 2017-18)\n\nTitle: “Society and Teenagers: How statistics reveal the changes in young peoples lives through the last century”\n\nAbstract: “Young people born in the year 2000 in the UK are more likely to be called Megan or Jack than Mary or John\, had a life expectancy at birth of over 75 years and about half will go on to higher education. How was this different a hundred years ago? Official statistics tell the story of how our lives have developed over a century of dramatic change. Statisticians have played a key role in tracking and understanding these developments. I will talk about the statistics behind the measurement of our daily lives. For example\, how a basket of goods is used to represent the way people spend their money and what the items in the basket tell us about changes in society and technology. The talk will also discuss the measurement of poverty and its development over time into more than just an indicator of having enough food to eat. I will include examples of what life was like for teenagers through the last century as revealed by the rich source of official statistics.
URL:http://math-sciences.org/event/society-and-teenagers-how-statistics-reveal-the-changes-in-young-peoples-lives-through-the-last-century
CATEGORIES:Statistics and Data Science
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=UTC+0:20170626T160000
DTEND;TZID=UTC+0:20170626T170000
DTSTAMP:20170624T051651
CREATED:20170608T110135Z
LAST-MODIFIED:20170615T125839Z
UID:2498-1498492800-1498496400@math-sciences.org
SUMMARY:Barry Dillon (Plymouth)
DESCRIPTION:Compositeness beyond the Standard Model\n\nOne of the leading motivations in the search for new physics at the LHC is the idea of `naturalness'. As it stands\, the current Standard Model (SM) of particle physics is deemed unnatural due the unexplained smallness of the Higgs boson mass\, and due to the unexplained hierarchical nature of the Yukawa couplings. A promising explanation of these features is that the SM Higgs is composite\, and that its fermions are partially-composite. Much like in QCD\, this model predicts a plethora of new composite states with masses near the confinement scale. However the absence of any new physics detection at the LHC puts stringent constraints on the current models. In this talk we will discuss the Minimal Composite Higgs Model (MCHM) and how 5D holographic techniques can be used as calculational tools to derive predictions. We will then present results from a study of the top-partner spectrum in the holographic MCHM (1510.08482)\, and phenomenological bounds on heavy spin-2 composite states using the latest LHC results (1603.09550).
URL:http://math-sciences.org/event/barry-dillon-plymouth
LOCATION:2-5 Kirkby Place\, Plymouth\, PL4 6DT\, United Kingdom
CATEGORIES:Seminars,Theoretical Physics
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=UTC+0:20170630T140000
DTEND;TZID=UTC+0:20170630T150000
DTSTAMP:20170624T051651
CREATED:20170616T132240Z
LAST-MODIFIED:20170616T132240Z
UID:2512-1498831200-1498834800@math-sciences.org
SUMMARY:Yang Han (University of Exeter)
DESCRIPTION:Speaker: Yang Han (University of Exeter)\n\nTitle: Simultaneous Confidence Bands in Linear Regression\n\n \n\nAbstract: Construction of simultaneous confidence bands for a percentile line has been considered by several authors. However only conservative symmetric bands\, which use critical constants over the whole covariate range are available in the literature. Our new methods allow the construction of exact symmetric bands for a percentile line over a finite interval of the covariate . The exact symmetric bands can be substantially narrower than the corresponding conservative symmetric bands. Several exact symmetric confidence bands are compared under the average band-width criterion. Furthermore\, new asymmetric confidence bands for a percentile line are proposed. They are uniformly (and can be very substantially) narrower than the corresponding exact symmetric bands. Therefore\, asymmetric bands should always be used under the average band-width criterion. The proposed methods are illustrated with a real example of a drug stability study.\n\nMany modern medicines are targeted therapies\, targeting specific pathways. A biomarker that is informative of how sick a patient in the targeted pathway is could be sufficiently predictive of the effect on the patients to allow such medicine to be personalized. Baseline HbA1c for diabetic patients is an example of such potential biomarkers. If a candidate biomarker is continuously valued\, it is typically dichotomized to classify patients into target (marker-positive) and non-target (marker-negative) subgroups. The question\, for each potential cut-point\, is therefore whether the drug has sufficient efficacy in the overall population\, or only in the marker-positive patients\, or neither. This question can be fully answered by providing simultaneous confidence intervals on the effect of the drug on the marker-positive patients\, on the marker-negative patients\, and on their mixture. To confidently decide whether a continuously-valued biomarker is useful for targeting patients\, such simultaneous confidence intervals need to be further adjusted for the multiplicity of searching through all possible cut-point values. This presentation gives\, for continuously-valued outcome measures such as reduction in HbA1c for Type II Diabetes\, a neat method providing exact (fully) multiplicity-adjusted simultaneously confidence intervals. An app is also available\, which conveniently implements the method we propose.
URL:http://math-sciences.org/event/yang-han-university-of-exeter
CATEGORIES:Statistics and Data Science
END:VEVENT
END:VCALENDAR