Department of Computing Seminars 1997/1998

Measuring Test Effectiveness - An Experiment on Operational Avionics Software
Tuesday October 28th 1997 - Stuart Reid, Cranfield:

A new approach to some classical results in Analysis and applications
Thursday Nov 6th 1997 - Professor Maria-Victoria Velasco, University of Granada:

A Graph-Theoretic Model of Human Memory
Tuesday 18th November 1997 - Professor Robin Whitty, Southbank University:
The modelling of memory is at once the province of psychologists, neurophysiologists and computer scientists. I shall try to give a flavour of what is going on in each domain and will offer a model of my own, a naive bit of graph theory which nevertheless seems to have some appealing features.

Random Necklaces, Graphs and Molecules: The Challenge of Unlabelled Sampling
Tuesday November 25th 1997 - Dr. Leslie Ann Goldberg, University of Warwick:
Informally, an "unlabelled combinatorial structure" is an object such as an unlabelled graph (in which the vertices are indistinguishable) or a structural isomer in chemistry (which shows how the various atoms of a molecule should be connected, without further distinguishing between different atoms of the same type). This talk is concerned with the question ``How do you sample uniformly at random from a set of unlabelled combinatorial structures?''. The talk will survey known algorithmic methods that apply to such problems, and will point out limitations of the known methods. Some of the many open problems in this area will be discussed.

Transparent Access To Distributed Information Sources
Wednesday December 3rd 1997 - Dr. Norman Paton, Manchester University:
The TAMBIS (Transparent Access to Multiple Biological Information Sources) project at Manchester University is developing a system that allows declarative queries to be expressed that range over highly diverse information sources. Users express queries using a visual interface that presents the illusion that there is only one information source, described using a rich conceptual model. This is a big lie! The seminar outlines the TAMBIS architecture, and indicates how the illusion is maintained that biological sources are much cleaner, more integrated and fewer in number than is actually the case.

Test adequacy criteria and their relations to fault detecting ability and test cost
Tuesday December 9th 1997 - Dr Hong Zhu, Institute of Computer Software, Nanjing University:

Constructions for cubic graphs with large girth
Tuesday January 20th 1998 - Professor Norman Biggs, LSE:
The aim of this talk is to give a coherent account of the problem of constructing cubic graphs with large girth. There is a well-defined integer $\mu_0(g)$, the smallest number of vertices for which a cubic graph with girth at least $g$ exists, and furthermore, the minimum value $\mu_0(g)$ is attained by a graph whose girth is exactly $g$. The values of $\mu_0(g)$ when $3 \le g \le 8$ have been known for over thirty years. For these values of $g$ each minimal graph is unique and, apart from the case $g=7$, a simple lower bound is attained. \smallskip The talk will describe what has been achieved for $g \ge 9$, where the situation is quite different. Here it is known that the simple lower bound is attained if and only if $g=12$. A number of techniques will be mentioned, with emphasis on the construction of families of graphs $\{ G_i\}$ for which the number of vertices $n_i$ and the girth $g_i$ are such that $n_i\le 2^{cg_i}$ for some finite constant $c$.

Predicting software project effort using machine learning technology
Tuesday February 10th 1998 - Professor Martin Shepperd, Bournemouth University:

Discretely observed multivariate continuous time autoregressive processes
Thursday February 12th 1998 - Georgi Boshnakov (Bulgaria, visiting Goldsmiths):

Application of Network Flows to Rearrangement of Series
Tuesday February 24th 1998 - Professor Nash-Williams, Reading University:
This describes joint work with my colleague David White on the following problem. For each permutation f of the set of positive integers, determine all triples s, t, u of real numbers such that t and u are the upper and lower limits of the sequence of partial sums of the "f-rearrangement" a(f(1)) + a(f(2)) + a(f(3)) + ... of some convergent real series a(1) + a(2) + a(3) + ... with sum s. Somewhat unexpectedly, the Max-Flow Min-Cut Theorem concerning network flows proves very useful in dealing with this question.

A Bayesian semiparametric accelerated failure time model and extensions
Thursday March 12th 1998 - Bani Mallick (Imperial College):

A latent trait model for a mixture of binary, polytomous, normal and gamma items.
Thursday March 19th 1998 - Irini Moustaki (LSE):
A general framework will be discussed for analyzing data with distributions from the exponential family using a latent trait model. Factor scores will be also discussed and put in a general framework. Two applications will be presented to illustrate the model. One is from the Eurobarometer Survey, 1992, on attitudes towards Science and Technology and the other is from the Enquete sur la consommation 1990 survey in Switzerland.

Radio Channel Assignment
Tuesday March 24th 1998 - Dr Robert Leese, Oxford University:

The automation of software testing using genetic algorithms
Tuesday, 28th April 1998 - Professor Bryan Jones, Glamorgan University:
Software testing may take up to 50% of the effort involved in developing software. Whilst many tools have been developed to automate regression testing, a tool that automatically generates test cases from the software specification or code remains elusive. Genetic algorithms have been used successfully to generate tests that cover all branches using the branch predicate as the basis of a fitness function. These ideas have been extended to find the longest and shortest execution times of real-time software and to search for commonly occurring programming faults (fault-based testing). The success achieved so far suggests that future developments may lead to a viable computer aided software testing tool.

Software Quality Management in practice - how effective is ISO 9000?
Tuesday May 12th 1998 - Professor Frank Bott, Aberystwyth University:
I have recently been involved in two major software package procurements, each costing between 100,000 and 150,000 pounds. This experience and the subsequent experience of chairing the implementation working groups leads me to believe that ISO 9000 is ineffective and that this is as much the fault of the customers as of the suppliers.

A partial order on permutations with applications to Computer Science
Thursday May 14th 1998 - Professor Michael Atkinson, University of St Andrews:
Some properties of permutations which constrain the order of the permuted symbols are inherited by subsequences. Examples of such properties in Computer Science and Combinatorics will be given, and the partial order which underlies the idea of "permutation inheritance" will be defined and studied.

Symmetry in graph labeling problems
Wednesday May 27th - Douglas Rogers.
In recent years, a wide variety of graph labeling problems have come to the fore, with many applications. These problems are often challenging, but in their nature seem to admit few general techniques. In this talk, we consider several of these problems in the case of the complete k-ary tree, by way of exploring the significance of symmetry in these problems. For example, we discuss an algorithm for gracefully labeling the complete k-ary tree; and we also examine a method of proof aimed at showing that the additive bandwidth of the complete k-ary tree is the same as the bandwidth.

Personalizable Hyperlink-based Systems
Tuesday July 14th - James Ohene-Djan.