Transcript PowerPoint at Reading - Council for the Mathematical Sciences

Mathematics for the Digital
Economy
Building Stones
Roland Potthast, Reading, UK
© University of Reading 2009
www.reading.ac.uk
The Digital Economy
“Novel design or use of information and communication technologies to help transform the lives of individuals,
society or business.” (EPSRC)
New
Technology
+
Behaviour
Lifestyle
=
Highly multidisciplinary: high impact
Technology + People/Society = Resonance = New
Opportunities
To put your footer here go to View > Header and Footer
2
Math and Digital Economy
Products
Team
People
Modelling
Area/Data
To put your footer here go to View > Header and Footer
Services
Algorithms
Structures
3
People
Example CV (Potth.) Bridging
the Gaps:
To put your footer here go to View > Header and Footer
-
Education Physics and
Mathematics
-
Project Manager in IT
Industry
-
Trainer for Siemens ICN
-
Partner for Spin-Off
Companies in IT/Maths
-
Mathematics
Reader/Professor
-
Team Builder
4
Math and Digital Economy
Products
Team
People
Modelling
Area/Data
To put your footer here go to View > Header and Footer
Services
Algorithms
Structures
5
Becoming data rich…
Data from many sources
–
–
–
–
–
–
–
–
Behaviour of people and groups
Transactions B2C
Communication/networking P2P
Outreach and information
Participation
Monitoring/surveillance
Measurement
New Imaging Technologies
• In many areas we are data rich but model
poor!
• Especially when the “atoms” are people
To put your footer here go to View > Header and Footer
6
Math and Digital Economy
Products
Team
People
Modelling
Area/Data
To put your footer here go to View > Header and Footer
Services
Algorithms
Structures
7
… the information revolution
The information age just started!
• More information and data on various levels than we
could ever imagine: economy, society, science
• Sincere demand of models, order, understanding,
monitoring, control
• Scaling, Micro vs. Macro Analysis,
• Hierarchy of Models,
• New Mathematics,
continuous or discrete!
To put your footer here go to View > Header and Footer
8
Math and Digital Economy
Companies
Products
Team
People
Modelling
Area/Data
To put your footer here go to View > Header and Footer
Services
Algorithms
Structures
9
Some Digital Players
Now part of the fabric of our lives
Online commerce
Pure plays
Existing
Finance
Social interaction
Networking
Commerce
Services
Communications
Utilities…
To put your footer here go to View > Header and Footer
10
Math and Digital Economy
Products
Team
People
Modelling
Area/Data
To put your footer here go to View > Header and Footer
Services
Algorithms
Structures
11
Customer Relationship Management
•
Explosive growth though IT “reach” … 105-106
customers
•
Using behaviour to discover/define addressable
groups
•
Highly responsive : near real time
•
Finding markets that are predictive
•
Predicting behaviour and churn
SORTING OUT THE CROWD
Maths Inside
Unsupervised discrimination over data bases: EM algorithm and its variants
Hidden Markov models for rates of transition between behavioural states
Supervised discrimination: Bayes factors and probability theory
Discrete searches model optimisation: genetic algorithms
Simulation : Agent Based Modelling e.g. possible spin out from Maths@UoR
To put your footer here go to View > Header and Footer
12
Cognitive Neuroscience / Healthcare
• Exploding number of new imaging technologies
• Aging society with new needs of diagnostics
• Societies growing strongly in developing countries
• Time-resolved multi-source data, need for model hierarchy and evaluation
Maths Inside
Discrete Theories and Field Theories, Integro-Differential Equations
Medical Imaging, Monitoring, Data Analysis, Remote Analysis
Automated algorithms, Remote Health Care
Inverse Problems, Data Assimilation, Stochastic Estimation Theory
New Multi-Level Structures, Models, Analysis and Numerics
To put your footer here go to View > Header and Footer
13
Monitoring and security
•
Searching for aberrant or low probability events
•
Classifying behaviour
•
Prioritising for different types of intervention
•
Supply Chain Management and Monitoring
SiroTechnologies
(SiroTechnologies, EADS, VW, RLS, KMW etc)
•
e.g. Health-check data in the home and online
(www.brainpanrel.co.uk bid to LLHW prog),
•
e.g. fraudulent behaviour detection for online poker companies (Valeo
Associates Ltd)
Maths Inside
Bayesian multiple hypotheses testing
Forecasting trends/uncertainties : application of MCMC in adaptive forecasts
Supervised discrimination: Log Bayes factors / probability theory
To put your footer here go to View > Header and Footer
14
Graphs and Networks
•
The growth and evolution of dynamical networks
•
Small world and range dependent graphs
– e.g.
•
Inverse problems: calibrating graph parameters from data
•
Dealing with very large networks – sensitivity to data
•
Comparison of alternative concepts/models
Maths Inside
New classes of random graphs
Numerical linear algebra & spectral theory: clustering within networks
Generalised clustering methods, e.g. SVD-based for stochastic graphs
Maximum likelihood representations of data within classes of graphs
Stability of results with respect to data, Inverse Problems for Graphs
To put your footer here go to View > Header and Footer
15
Behaviour based profiling segmentation
•
Segmenting populations with behavioural metrics for product and
service development
•
e.g. Smart 24/7 energy metering data in the home – “current insight”
mining pilot project
•
e.g. Analysing m-banking data in Africa (current 2M customer pilot
UoOx start-up ARK MF Ltd)
Maths Inside
Unsupervised discrimination over data bases: EM algorithm and its variants
Markov models for rates of transition between behavioural states
Supervised discrimination: Bayes factors and probability theory
Discrete search algorithms: genetic algorithms
Simulation : Agent Based Modelling
To put your footer here go to View > Header and Footer
16
Fuel Cell Quality Monitoring
•
Monitoring of current distributions in fuel cells via
magnetic tomography
•
Development of fuel cells and fuel cell stacks, energy
patterns, applications
•
Production and quality control
•
Maintenance, Diagnostics, Control
Maths Inside
Integral equations and Potential Theory, PDE, Numerical Analysis
Inverse Problems, Imaging, Data Assimilation, Optimization Algorithms
Data analysis, Large ODE systems, FEM/FIT/BEM
Unsupervised discrimination over data bases: EM algorithm and its variants
To put your footer here go to View > Header and Footer
17
Math and Digital Economy
KT
Products
Team
People
Modelling
Area/Data
To put your footer here go to View > Header and Footer
Services
Algorithms
Structures
18
KT Opportunities
• Many companies have these topics
• Need for new concepts and new practical
applications
• Data is very often confidential which is a
barrier
• The maths community would benefit
from anonymous problem banks
• Algorithms and methods are difficult to
protect
– Secrecy rather than publication
To put your footer here go to View > Header and Footer
19
The Horizon Hub
•
•
University of Nottingham and
“spokes” at Reading, Cambridge,
Exeter
Highly multidisciplinary:
–
–
–
–
–
•
•
•
ICT,
Maths,
Business,
Social science
Art, Performance
Starts is a few months
Large and growing number of
industrial partners
Grindrod, UoR, will manage an
interface with UK advertising
companies on behalf of the national
DE community – The IPA
To put your footer here go to View > Header and Footer
20
Math and Digital Economy:
Future Trends
•
Integration of diverse and multilevel technologies
•
Simplification and complexity
•
Individual empowerment: customers’ perceptions and activities and ideas becoming
paramount
•
Commerce P2P exchange
•
Control, Security, Sustainability
•
Ethics and individual/subjective issues
Products
Team
People
Modelling
Area/Data
Services
Algorithms
Structures
To put your footer here go to View > Header and Footer
Thank you!
21