DEPARTMENT OF COMPUTING SCIENCE AND MATHEMATICS
UNIVERSITY . COMPUTING SCIENCE . SEMINARS

SEMINARS - Autumn 2009

[Talk Schedule] [Abstracts] [Previous Seminars]

The Department of Computing Science and Mathematics presents the following seminars. Unless otherwise stated, seminars will take place in Room 4B94 of the Cottrell Building, University of Stirling from 15.00 to 16.00 on Friday afternoons during semester time. For instructions on how to get to the University, please look at the following routes.

If you would like to give a seminar to the department in future or if you need more information, please contact the seminar organiser, Savi Maharaj (Phone 01786 467431, Email savi@cs.stir.ac.uk).

Talk Schedule [Top] [Abstracts]

Abstracts [Top] [Schedule]

25th September [Schedule]

Modelling mechanisms of legged locomotion: Brain controlled gait pattern generation
Zhijun Yang
School of Engineering
The University of Edinburgh

Abstract

As an engine of almost all life phenomena, neural oscillation networks play a crucial role in the activities of all animals. In this seminar, a novel, model-independent approach is introduced for the retrieval of coupled neural oscillations observed in biological CPGs during the control of walking. Based on a simple distributed synchroniser, various Oscillatory Building Blocks (OBBs), which operates in discrete time domain and amenable to digital circuit implementation, can be configured for the production of complicated rhythmic patterns. A general method is provided for the construction of a target artificial CPGs architecture behaving as an asymmetric Hopfield network. In addition, a new analog VLSI implementation of a simple OBB module is also presented which can operate in the continuous time domain.

2nd October [Schedule]

Research in a Connected World: using advanced IT to empower researchers
Alex Voss
Edinburgh e-Science Centre & University of St Andrews

Abstract

Research today is often critically dependent on computation and data handling. Enabled by worldwide networks, sophisticated analytical methods and research data archives, researchers today have access to a wide array of data about the natural world, about societies and about engineered structures. Powerful simulation methods can be used to project existing data forward in time or to apply it to hypothetical systems and societal structures. The availability of data in many areas of research outstrips our ability to analyse it effectively, to draw conclusions and generate knowledge.This talk will present some of the ways in which researchers harness the power of advanced networked information technologies to develop e-Infrastructures for research that help to address research challenges of global importance. It will discuss the problem of developing research tools that are not only fit for purpose but that enable researchers from other disciplines to conduct cutting-edge research. Developing systems that embody novel research methods is an exciting but challenging task.

9th October [Schedule]

Role of Domain Engineering in Software Development - Why Current Requirements Engineering is Flawed
Dines Bjorner
The University of Edinburgh (SICSA visitor)

Abstract

We introduce the notion of domain descriptions (D) in order to ensure that software (S) is right and is the right software, that is, that it is correct with respect to written requirements (R) and that it meets customer expectations (D). That is, before software can be designed (S) we must make sure we understand the requirements (R), and before we can express the requirements we must make sure that we understand the application domain (D): the area of activity of the users of the required software, before and after installment of such software. We shall outline what we mean by informal, narrative and formal domain descriptions, and how one can systematically -- albeit not (in fact: never) automatically -- go from domain descriptions to requirements prescriptions. As it seems that domain engineering is a relatively new discipline within software engineering we shall mostly focus on domain engineering and discuss its necessity. The talk will show some formulas but they are really not meant to be read, let alone understood. They are merely there to bring home the point: Professional software engineering, like other professional engineering branches rely on and use mathematics. And it is all very simple to learn and practise anyway! We end this talk with, to some, perhaps, controversial remarks: Requirements engineering, as pursued today, researched, taught and practised, is outdated, is thus fundamentally flawed. We shall justify this claim.

16th October [Schedule]

Plastic Machines: Behavioural Diversity and the Turing Test
Michael Wheeler
Department of Philosophy
University of Stirling

Abstract

After proposing his famous Turing Test, Alan Turing himself considered a number of objections to the idea that a machine might eventually pass it. One of the objections discussed by Turing was that no machine will ever pass the Turing Test because no machine will ever "have as much diversity of behaviour as a man". He responded as follows: the "criticism that a machine cannot have much diversity of behaviour is just a way of saying that it cannot have much storage capacity". I shall argue that the objection cannot be dismissed so easily. The diversity exhibited by human behaviour is characterized by a kind of context-sensitive adaptive plasticity. Most of the time, human beings flexibly and fluently respond to what is relevant in a given situation. Moreover, ordinary human life involves an open-ended flow of shifting contexts to which our behaviour typically adapts in real time. For a machine to "have as much diversity of behaviour as a man" would be for that machine to keep its responses and behaviour relevant within such a flow. Merely giving a machine the capacity to store a huge amount of information and an enormous number of behaviour-generating rules will not achieve this goal. By drawing on arguments presented originally by Descartes, by making contact with the frame problem in artificial intelligence, and by touching base with recent work in evolutionary robotics, I shall argue that the distinctive context-sensitive adaptive plasticity of human behaviour explains why the Turing Test is such a stringent test for the presence of thought, and why it is much harder to pass than Turing himself may have realized.

23rd October [Schedule]

Goals and Policies for Telephony
Ken Turner
Department of Computing Science and Mathematics
University of Stirling

Abstract

Goals are abstract application-oriented objectives for how a system should behave. To be made operational, they are refined into lower-level policies that are executed dynamically. This talk will explain the design of ACCENT (Advanced Component Control Enhancing Network Technologies, http://www.cs.stir.ac.uk/accent) - an automated system for support of goals and policies. The approach will be explained with respect to Internet telephony, although it has also been used to manage other kinds of systems such as sensor networks and home networks.

 

4th November [Schedule]

Climate Modelling: A beginners guide
Simon Tett
Chair of Earth System Dynamics
School of Geosciences
The University of Edinburgh

Abstract

This talk will explain how, in broad terms, climate models work. It will explain where and why uncertainties arise and what those mean for future projections of climate. It will also outline why climate researchers believe it is likely than humans are responsible for the warming observed since the mid 20th century.

13th November [Schedule]

Software Engineering at Adobe
James Boag, Ian Paterson and Adrian O'Lenskie
Adobe, Edinburgh

Abstract

Adobe is a well-established, world-wide software producer. With a wide range of products and technologies, they are an extremely well known brand (even my granny has heard of them). Three employees from their Edinburgh office plan to give some insights into working at Adobe, and the industry in general.

James Boag is a relatively new addition to the team, so new he is still called "young James". James will describe what it was like joining Adobe, lessons he's learned since graduating, and promises never to mention regrets he has.

Ian Paterson is a seasoned Computer Scientist at Adobe and is the technical lead for the Edinburgh group. He is sought after within the company for technical guidance and is constructing Web 3.0 technology to manage his fishing club. Ian will talk about the qualities that successful engineers share. He has been warned not to mention fishing.

Imagine the genetically engineered love-child of the pointy haired boss in Dilbert and David Brent. You'd be thinking of Adrian O'Lenskie. An engineering manager, Adrian was formerly a Computer Scientist working on products such as Adobe InDesign and Adobe Acrobat. He was found out and now spends his time corralling the smart folks. Adrian will talk about some of the typical (and atypical) team structures found at Adobe and how, from a corporate perspective, we "get the job done".

20th November [Schedule]

Young Children Learning with Toys and Technology at Home
Lydia Plowman & Olivia Stevenson
Stirling Institute of Education

Abstract

The aims of Young Children Learning with Toys and Technology at Home (funded by the ESRC, 2008 - 2011) are i) to use household case studies to produce a richly detailed account of young children's encounters with technology in the home and ii) to extend methods for examining children's experiences of technology in their domestic environments. Our visits over a period of a year or so to three- and four-year-old children and their families in fourteen households use conversations, video, and photographs to describe interactions with a range of domestic, leisure and work technologies. As we complete data collection, we reflect on the research to date and present case studies of some technological toys.

 

27th November [Schedule]

Riemannian Manifold Hamiltonian Monte Carlo
Mark Girolami
Department of Computing Science
University of Glasgow

Abstract

This talk will present a Riemannian Manifold Hamiltonian Monte Carlo sampler to resolve the shortcomings of existing Monte Carlo algorithms when sampling from target densities that may be high dimensional and exhibit strong correlations. The method provides a fully automated adaptation mechanism that circumvents the costly pilot runs required to tune proposal densities for Metropolis-Hastings or indeed Hybrid Monte Carlo and Metropolis Adjusted Langevin Algorithms. This allows for highly efficient sampling even in very high dimensions where different scalings may be required for the transient and stationary phases of the Markov chain. The proposed method exploits the Riemannian structure of the parameter space of statistical models and thus automatically adapts to the local manifold structure at each step based on the metric tensor. A semi-explicit second order symplectic integrator for non-separable Hamiltonians is derived for simulating paths across this manifold which provides highly efficient convergence and exploration of the target density. The performance of the Riemannian Manifold Hamiltonian Monte Carlo method is assessed by performing posterior inference on logistic regression models, log-Gaussian Cox point processes, stochastic volatility models, and Bayesian estimation of parameter posteriors of dynamical systems described by nonlinear differential equations. Substantial improvements in the time normalised Effective Sample Size are reported when compared to alternative sampling approaches. Matlab code at http://www.dcs.gla.ac.uk/inference/rmhmc allows replication of all results.

Previous Seminar Series [Top] [Abstracts] [Schedule]

Last Modified: 24th November 2009