Speakers
Koichiro Akiyama (Toshiba Corporation, Japan)
A Survey of Algebraic Surface Cryptosystems
Algebraic surface cryptosystems(ASC) is a public-key cryptosystem
whose security is based on a hard problem in algebraic geometry. ASC
key size is one of the shortest among those of post-quantum public-key
cryptosystems known at present. Several ASC was proposed until now.
In this talk, I will show you proposed ASC and attacks on that.
Noel Barton (Sunoba Pty Ltd., Australia)
Mathematical Simulation Answers Investor Questions on New Concepts for Solar Power
The speaker has worked since 2004 on a new solar heat engine. Potential investors want a measure of certainty before they will invest, and the lecture will describe how some required information was created. The analysis to be presented includes a thermodynamic model of the engine, as well as a simulation of heat storage and power generation at a suitable site.
Andreas Binder (MathConsult GmbH, Austria)
Model Uncertainty in Computational Finance
In the valuation and risk management of structured financial instruments,
various models for the stochastic movement of the underlyings (e.q.
equities, interest rates, exchange rates) are used.
We present efficient numerical schemes to calibrate model parameters to
given market data of liquid instruments. A good quality of these fits does
not assure that calculated prices for more complicated instruments also lie
within a tight price band for various stochastic models.
Hayato Chiba (IMI, Kyushu University, Japan)
A spectral theory of linear operators on Gelfand triplets and its applications to infinite dimensional dynamical systems
Synchronization phenomena in a system of large populations of coupled oscillators have been of great interest. The Kuramoto model is often used to investigate such phenomena.
In this talk, an infinite dimensional Kuramoto model is considered, and the Kuramoto's conjecture on a bifurcation diagram of the system, which is open since 1985, is proved.
For this purpose, a new spectral theory of linear operators based on Gelfand triplets is developed. Basic notions in the usual spectral theory, such as eigenspaces,
algebraic multiplicities, point/continuous/residual spectra, Riesz projections are extended to those defined on a Gelfand triplet.
By using them, bifurcations in an infinite-dim space will be treated.
Xavier Dahan (FM, Kyushu University, Japan)
Application of height theory to some modular algorithms in Symbolic Computation
"Accurate" and "reliable" computations involve by essence a large amount of data.
Numerical approximations do not always guarantee such precise, reliable results.
On the opther hand, computing exactly, as relevant of symbolic computation
involve often very large numbers, and even larger during intermediate computations.
It is the case for instance when dealing with polynomials with several unknowns.
There is a classical recipe, called "modular methods",
to avoid coefficients to become too large.
This recipe raises several questions, and an important one concerns complexity.
I will show how a tool coming from Diophantine geometry, called heights, can
help to give an answer to this question.
Kathy Horadam (RMIT University, Australia)
Information Networks, Biometrics and Points Between
This talk will describe recent research in two very diverse areas:
improving biometric matching of ridge or vessel based biometrics, and
finding locally interesting communities in complex networks. We have
found some surprising common ideas. I will also include some lessons
learned working in such applied areas when coming from a pure
mathematics background.
Jonathan Hosking (IBM T. J. Watson Research Center, U.S.A.)
Short-Term Forecasting of the Daily Load Curve for Residential Electricity Usage in the Smart Grid
We describe a model of the daily load curve for residential electricity
consumption that includes the effects of dynamic price incentives on
demand response. This linear Gaussian state-space model provides
estimates of the intraday load substitution effects that are induced by
the pricing schedules, and enables fast updating of load forecasts as
new usage and weather data arrive.
Alejandro Jofré (CMM and DIM, Universidad de Chile, Chile)
Optimization and Game Theory in Energy Markets
In this presentation we explain how stochastic optimization and game theory techniques are relevant to understand energy markets today. Issues related to operation and regulation of a generation/transmission network system represented by a graph in which generators are located at nodes and the arcs are transmission lines are analyzed. Different pricing rules are deduced from a principal agent model, asymmetric information, regulation and market power are analyzed. We also develop simulations to visualize social optimum versus current pricing rules in this framework.
Naoyuki Kamiyama (IMI, Kyushu University, Japan)
Pareto Stability in a Two-sided Matching Market with Indifferences
Since the seminal work of Gale and Shapley, two-sided matching markets
have been extensively studied in both Economics and Computer Science.
In this model, there exist two groups of agents and each agent has a preference
ranking over members of the other group. The goal is to find a matching between
these two groups with some specified properties. In two-sided matching markets,
the concept of stability proposed by Gale and Shapley is one of the most
important solution concepts. However, it is known that if we introduce
ties into
preference lists, then stability do not guarantee Pareto efficiency
that is also
one of the most important solution concepts in two-sided matching markets.
This fact naturally leads to the concept of Pareto stability, i.e.,
both stable and
Pareto efficient. In this talk, we give a survey on recent developments of
this new solution concept.
Sei Kato (IBM Global Business Services, Japan)
Big Data, Big Simulation, and Deep Learning
Recent progress in distributed processing technologies has opened the door to a new world of massive computing. One example is big data. To date, no practical means has existed to gain insights from ever-growing accumulations of data, but the emergence of distributed streaming computing platforms has realized the continuous analysis of big data, which is crucial for continual business use. Regarding big simulations, a large-scale traffic flow simulation on a peta-scale computer has been realized by the development of an X10-based ultra-large agent-based simulation platform on massively parallel supercomputers. A QA system for diagnostic assistance is now becoming a reality based on IBM's Watson, which has been realized partly through its incorporation of deep learning technologies. This talk presents a review of state-of-the art computer platforms for massive computing, in addition to application examples.
Atsushi Kawamoto (TOYOTA CENTRAL R&D LABS., INC., Japan)
Topology Optimization in Multiphysics Problems
This lecture explains the basic concept of topology optimization,
discusses the involved methodological elements, and focuses on some
specific issues to multiphysics problems. Also, our research activities
in this field are partly presented such as thermal stress relaxation of
power semiconductor devices, designing heat exchangers and prototyping
lightweight car seat structures.
Frank Lutz (Technische Universität Berlin, Germany)
Discrete Topology of Grains and Periodic Foams
Our aim is to use methods from discrete and geometric topology
to recover structural information from the composition
of polycrystalline materials (such as metals and certain
ceramics) and also of monocrystalline materials that have
a periodic foam structure (such as gas hydrates and
transition metal alloys).
Osamu Maruyama (IMI, Kyushu University, Japan)
Protein Complex Prediction
In this talk, we will consider the problem of protein complex
prediction, which is a challenging problem in computational biology.
After a brief introduction of this problem, we will consider a novel
method based on an MCMC (Markov chain Monte Carlo) sampling method.
Robert Mckibbin (Massey University, New Zealand)
Industrial Mathematics Adventures in the South-West Pacific
A very, if not the most, important part of gathering information in industrial mathematics is that of the "interdisciplinary conversations" that take place between the industry people and the mathematicians. The mathematicians and statisticians can only help to construct good quantitative models when they fully understand the problems while the industry people can only take advantage of the mathematicians' skills when they realise the possibilities. This talk will describe some "conversations" that have taken place during the "information recovery and discovery" stage of conceptual modelling in ANZIAM Mathematics-in-Industry groups over the last few years.
Yoshihiro Mizoguchi (IMI, Kyushu University, Japan)
Mathematical Aspects of Interpolation Technique for Computer Graphics
We introduce a simple mathematical framework for 2D shape interpolation
methods that preserve rigidity. Several existing rigid shape interpolation
techniques are discussed and mathematically analyzed through our framework.
We also introduce the Laplacian Matrix of a graph and show some several
properties about eigenvalues and eigenvectors in connection with image
segmentations and group formations.
Junichi Nakagawa (Nippon Steel ＆ Sumitomo Metal Corporation, Japan)
Mathematical Core Technology for a Traffic Flow System
Using the Model Predictive Control of Large and Complex Networks
Traffic jams have caused a lot of adverse effects on our society regarding fuel consumption, exhaust gas, physical distribution, and so on.
An important theme of this project is the establishment of mathematical core technologies regarding traffic control. We thus study the real-time optimization of large-scale and complex systems along with controls for multi-scalability between micro-scale objects that correspond to vehicles and macro-scale ones that correspond to traffic flow.
Junji Nakano (The Institute of Statistical Mathematics, Japan)
Some Topics in Computational Statistics and
Statistical Training
Computational statistics has grown with the developments of
computer technologies and now becomes the main stream of
statistics. I briefly explain some topics in it, including
statistical software and data visualization. Although such
modern statistical knowledge is strongly required now,
statistical training system in Japan has not been well
equipped for it. I explain our project to tackle this
situation, the school of statistical thinking at the ISM.
Yoshiyuki Ninomiya (IMI, Kyushu University, Japan)
Model Selection for Irregular Statistical Models
Irregular statistical models such as signal models, mixture models,
change-point models, hidden Markov models or factor models have been
widely used in industrial area, but their irregularity is sometimes
ignored. For example, naive information criteria do not work well
because conventional statistical asymptotic theories cannot be used
owing to the irregularity. In this presentation, the detail about
this problem will be explained, and a reasonable information criterion
will be proposed in some cases.
Konrad Polthier (Freie Universität Berlin, Germany)
Applications of Integrable Discrete Geometric Structures
In this presentation we introduce novel integrable discretization schemes for discrete
surface and volumetric meshes, and we stress their eciency in several industrial applications from scienti c computing, computer aided design, architecture and computer
graphics. Meshes arising from 3D scans or other 3D imaging techniques are often not well
adjusted to the geometry or topology of the underlying shapes. The generation of good
meshes is still an active research area in various disciplines. We discuss novel techniques
to generate highly structured surface and volume meshes that are consistent with many
geometric and topological features of the underlying shapes. We discuss techniques to
ll a bounded volumetric shape with a consistent cubical voxel structure. Among
the optimization goals are alignment of the voxels with the bounding surface as well
as simplicity of the voxel grid. Mathematical analysis of the possible singularities is
given. This "CubeCover" algorithm uses a tetrahedral volume mesh plus a user given
guiding frame eld as input. Then it constructs an atlas of chart functions, i.e. the
parameterization function of the volume, such that the images of the coordinate lines
align with the given frame eld. Formally the function is given by a rst-order PDE,
namely the gradients of the coordinate functions are the vectors of the frame. In a rst
step, the algorithm uses a discrete Hodge decomposition to assure local integrability
of the frame eld. A subsequent step assures global integrability along generators of
the rst homology group and alignment a face of the boundary cube with the original
surface boundary. All steps can be merged into solving linear equations.
Wayne Rossman (Kobe University, Japan)
Comparison of Smooth, Discrete and Semi-discrete Surfaces
Surfaces in differential geometry have been regarded classically
as mathematically-pure smooth objects. However, a recent trend
has been to consider discretized versions of surfaces, which is
highly useful in some applications (computer graphics and
architectural structure design, to name two). Our goal is to
understand how to discretize surfaces without discarding
advantageous differential geometric properties.
Zuowei Shen (National University of Singapore)
MRA based Wavelet Frame and Applications
One of the major driving forces in the area of applied and computational harmonic analysis during the last two decades is the development and the analysis of redundant systems that produce sparse approximations for classes of functions of interest. Such redundant systems include wavelet frames, ridgelets, curvelets and shearlets, to name a few. This talk focuses on tight wavelet frames that are derived from multiresolution analysis and their applications in imaging. The pillar of this theory is the unitary extension principle and its various generalizations, hence we will first give a brief survey on the development of extension principles. The extension principles allow for systematic constructions of wavelet frames that can be tailored to, and effectively used in, various problems in imaging science. We will discuss some of these applications of wavelet frames. The discussion will include frame-based image analysis and restorations, image inpainting, image denoising, image deblurring and blind deblurring, image decomposition, segmentation and CT image reconstruction.
Taizo Shirai (Sony Corporation, Japan)
Mathematical Approaches Used in Recent Cryptographic Technologies
Cryptographic technologies are key technologies to achieve security of the information society. It is required that the cryptographic technologies be withstand any types of attacks trying to bleach the protection. To evaluate the strength of cryptographic technologies, mathematics are used as very important tools to show guaranteed levels of security. In this talk, by looking at recent trend in cryptographic design of symmetric-key type technologies and public-key type technologies, we show some mathematical method used in them.
Jos Stam (Autodesk Research, Canada)
Fluid Dynamics for Entertainment
In this talk I will describe my research in computational fluid dynamics targeted to the entertainment industry such as film and games. I will cover the history of Fluid Dynamics and how these equations are solved in practice on a computer. The emphasis is on simple, stable and rapid solutions which yield visually realistic flows. The solver has been used in our MAYA Autodesk animation software and has received a Technical Achievement Award from the Academy of Motion Picture Arts and Sciences in 2007. We will also mention an iOS App called FluidFX that runs on iPhones and iPads based on the same technology. The talk will be accompanied by many live demonstrations and movies.
Tamon Suwa (FUJITSU LIMITED, Japan)
A Tsunami Simulator: Combination of the Particle-based Method and the Nonlinear Wave Model
We develop a tsunami simulator integrating a 3D fluid simulation technology using smoothed-particle hydrodynamics method, together with a 2D tsunami-propagation simulation technique using a nonlinear wave model.
The simulation will be performed on the K computer.
The numerical scheme of the simulator and examples of tsunami simulations will be reported.
We seek to contribute to improved disaster preparedness and disaster mitigation through a better understanding of a tsunami's mechanisms.
Daisuke Tagami (IMI, Kyushu University, Japan)
Efforts to Establish a Disaster Assessment
with the High Performance Computings
Still fresh in our memories is the huge earthquake in the Tohoku
area on March 11, 2011, where have been the worst affected by it.
To minimize damage, we need to figure out much complicated and
different scale phenomena during and after the earthquake. To
realize this object, we now try to establish a disaster assessment
with the high performance computings. In this talk, we introduce
our efforts; a fluid force assessment with the ISPH method and
a parallel computation of elaststatic problems with BDD-DIAG.
Koichiro Takayama (FUJITSU LABORATORIES LTD., Japan)
A Survey of Formal Verification Techniques for
Hardware Design and Expectations for Mathematics
Jun'ichi Takeuchi (Kyushu University, Japan)
Sparse Learning with Two Applications
Sparse learning or compressed sensing has been extensively
investigated for a decade in the field of information
science and engineering. We give a brief review on it,
following [Tanaka 2010]. Then, we discuss as its
applications the sparse superposition codes proposed by
Barron & Joseph (2010) in channel coding
for digital communications and Yang's super-resolution [Yang
et al 2008] in image processing.
Paul Urbach (Delft University of Technology, The Netherlands)
Mathematical Optimization for Applied Optical Problems
Optics is a crucial discipline in modern industry. This has been acknowledged by the European Union by giving photonics the status of Key-Enabling Technology (KET). In Optics mathematical optimization is used extensively and in widely different fields. Traditionally optimization methods are applied in classical optical design for complicated imaging systems such a in photolithography for integrated circuits. But optimization is also important in nano-optics, for example to design diffractive structures to couple light into a solar cell or to couple light out of a Led or to obtain highly sensitive plasmonic sensors. In optics many inverse problems are studied, for example in ellipsometry and in scatterometry. These inverse problems are in general also formulated as constrained optimization problems.
In most applications the number of optimization variables is finite, but sometimes it is easier to allow for infinite dimensional spaces in which optimum solutions are sought.
We shall give several examples of using mathematical optimization techniques to modern problems of applied optics.
Hayato Waki (IMI, Kyushu University, Japan)
Ill-conditionedness in Semidefinite Programming
We can obtain a tighter lower bound for a polynomial optimization
problem (POP) by using semidefinite programming (SDP) problems. For
some POPs, we need to solve very degenerate SDP problems. In general,
it is difficult to solve degenerate optimization problems accurately
even if such problems have convexity. However, we often obtain the
exact global optimal values of the original problems by means of
degeneracy in the SDP relaxation problems. In this talk, we present
such examples and talk about the reason why we can obtain the desired
values by using the degeneracy.
Masahiro Yamamoto (The University of Tokyo, Japan)
Industrial Mathematics by Inverse Problems: Expectations for Innovation by Mathematics
In industries, for effective operations of plants, optimal designs, etc., it is important how to interpret a lot of raw data.
The raw data do not directly describe the phenomena in the manufacturing processes under consideration and one has to extract the essence from raw data in quantitative ways. This is the inverse problem in the industrial mathematics. By relevant thinking by inverse problems, one can greatly improve recognition abilities of the processes, so that one can establish new guidelines for the processes, which yield great economic effects. I will present examples of inverse problems which are expected to yield innovations by mathematics.
Moti Yung (Google Inc., U.S.A.)
Cryptography: From Mathematics to Industrial Solutions
Cryptography is the old science of secret communication, and modern cryptograaphy also deals with integrity and secure remote transactions.
In this talk I will review the major characteristics of cryptography as
a mathematical and theoretical research discipline. I will then describe
how the principles of cryptography applies to the actual deployment of
cryptography in securing information technology systems, and I will present an example of a large scale system I was involved in.
