Report on BCTCS 2005

This report contains edited abstracts from BCTCS 2005, which was held on 22nd to 24th March 2005 in Nottingham, England.

Accurate Step Counting

Starting with an evaluator for a language, an abstract machine for the same language can be mechanically derived using successive program transformations. This has relevance to studying both the space and time properties of programs because these can be estimated by counting transitions of the abstract machine and measuring the size of the additional data structures needed, such as environments and stacks. In this article we use this process to derive a function that accurately counts the number of steps required to evaluate expressions in a simple language.

Journal publishing with Acrobat: the CAJUN project

The publication of material in electronic form should ideally preserve, in a unified document representation, all of the richness of the printed document while maintaining enough of its underlying structure to enable searching and other forms of semantic processing. Until recently it has been hard to find a document representation which combined these attributes and which also stood some chance of becoming a de facto multi-platform standard. This paper sets out experience gained within the Electronic Publishing Research Group at the University of Nottingham in using Adobe Acrobat software and its underlying PDF (Portable Document Format) notation. The CAJUN project1 (CD-ROM Acrobat Journals Using Networks) began in 1993 and has used Acrobat software to produce electronic versions of journal papers for network and CD-ROM dissemination. The paper describes the project's progress so far and also gives a brief assessment of PDF's suitability as a universal document interchange standard.

Separate compilation of structured documents

This paper draws a parallel between document preparation and the traditional processes of compilation and link editing for computer programs. A block-based document model is described which allows for separate compilation of various portions of a document. These portions are brought together and merged by a linker program, called dlink, whose pilot implementation is based on ditroff and on its underlying intermediate code. In the light of experiences with dlink the requirements for a universal object-module language for documents are discussed. These requirements often resemble the characteristics of the intermediate codes used by programming-language compilers but with interesting extra constraints which arise from the way documents are executed .

Adobe's Acrobat -- the Electronic Journal Catalyst?

Adobe's Acrobat software, released in June 1993, is based around a new Portable Document Format (PDF) which offers the possibility of being able to view and exchange electronic documents, independent of the originating software, across a wide variety of supported hardware platforms (PC, Macintosh, Sun UNIX etc.). The principal features of Acrobat are reviewed and its importance for libraries discussed in the context of experience already gained from the CAJUN project (CD-ROM Acrobat Journals Using Networks). This two-year project, funded by two well-known journal publishers, is investigating the use of Acrobat software for the electronic dissemination of journals, on CD-ROM and over networks.

Designing the Spectator Experience

Interaction is increasingly a public affair, taking place in our theatres, galleries, museums, exhibitions and on the city streets. This raises a new design challenge for HCI, questioning how a performer s interaction with a computer experienced is by spectators. We examine examples from art, performance and exhibition design, comparing them according to the extent to which they hide, partially reveal, transform, reveal or even amplify a performerts manipulations. We also examine the effects of these manipulations including movements, gestures and utterances that take place around direct input and output. This comparison reveals four broad design strategies: `secretive,' where manipulations and effects are largely hidden; `expressive,' where they are revealed, enabling the spectator to fully appreciate the performer's interaction; `magical,' where effects are revealed but the manipulations that caused them are hidden; and finally `suspenseful,' where manipulations are apparent, but effects only get revealed when the spectator takes their turn.

A Pyramidal Genetic Algorithm for Multiple-Choice Problems

This paper combines the idea of a hierarchical distributed genetic algorithm with different inter-agent partnering strategies. Cascading clusters of sub-populations are built from bottom up, with higher-level sub-populations optimising larger parts of the problem. Hence, higher-level sub-populations search a larger search space with a lower resolution whilst lower-level sub-populations search a smaller search space with a higher resolution. The effects of different partner selection schemes amongst the agents on solution quality are examined for two multiple-choice optimisation problems. It is shown that partnering strategies that exploit problem-specific knowledge are superior and can counter inappropriate (sub-) fitness measurements.

Partnering Strategies for Fitness Evaluation in a Pyramidal Evolutionary Algorithm

This paper combines the idea of a hierarchical distributed genetic algorithm with different inter-agent partnering strategies. Cascading clusters of sub-populations are built from bottom up, with higher-level sub-populations optimising larger parts of the problem. Hence higher-level sub-populations search a larger search space with a lower resolution whilst lower-level sub-populations search a smaller search space with a higher resolution. The effects of different partner selection schemes for (sub-)fitness evaluation purposes are examined for two multiple-choice optimisation problems. It is shown that random partnering strategies perform best by providing better sampling and more diversity.

The Danger Theory and Its Application to Artificial Immune Systems

Over the last decade, a new idea challenging the classical self-non-self viewpoint has become popular amongst immunologists. It is called the Danger Theory. In this conceptual paper, we look at this theory from the perspective of Artificial Immune System practitioners. An overview of the Danger Theory is presented with particular emphasis on analogies in the Artificial Immune Systems world. A number of potential application areas are then used to provide a framing for a critical assessment of the concept, and its relevance for Artificial Immune Systems.

A Recommender System based on the Immune Network

Abstract-The immune system is a complex biological system with a highly distributed, adaptive and self-organising nature. This paper presents an artificial immune system (AIS) that exploits some of these characteristics and is applied to the task of film recommendation by collaborative filtering (CF). Natural evolution and in particular the immune system have not been designed for classical optimisation. However, for this problem, we are not interested in finding a single optimum. Rather we intend to identify a sub-set of good matches on which recommendations can be based. It is our hypothesis that an AIS built on two central aspects of the biological immune system will be an ideal candidate to achieve this: Antigen - antibody interaction for matching and antibody - antibody interaction for diversity. Computational results are presented in support of this conjecture and compared to those found by other CF techniques.

'On the Effects of Idiotypic Interactions for Recommendation Communities in Artificial Immune Systems'

It has previously been shown that a recommender based on immune system idiotypic principles can outperform one based on correlation alone. This paper reports the results of work in progress, where we undertake some investigations into the nature of this beneficial effect. The initial findings are that the immune system recommender tends to produce different neighbourhoods, and that the superior performance of this recommender is due partly to the different neighbourhoods, and partly to the way that the idiotypic effect is used to weight each neighbour's recommendations.

'An Artificial Immune System as a Recommender System for Web Sites'

Artificial Immune Systems have been used successfully to build recommender systems for film databases. In this research, an attempt is made to extend this idea to web site recommendation. A collection of more than 1000 individuals' web profiles (alternatively called preferences / favourites / bookmarks file) will be used. URLs will be classified using the DMOZ (Directory Mozilla) database of the Open Directory Project as our ontology. This will then be used as the data for the Artificial Immune Systems rather than the actual addresses. The first attempt will involve using a simple classification code number coupled with the number of pages within that classification code. However, this implementation does not make use of the hierarchical tree-like structure of DMOZ. Consideration will then be given to the construction of a similarity measure for web profiles that makes use of this hierarchical information to build a better-informed Artificial Immune System.

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

'A Bayesian Optimisation Algorithm for the Nurse Scheduling Problem'

Abstract- A Bayesian optimization algorithm for the nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. Unlike our previous work that used GAs to implement implicit learning, the learning in the proposed algorithm is explicit, i.e. eventually, we will be able to identify and mix building blocks directly. The Bayesian optimization algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, a new rule string has been obtained. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed approach might be suitable for other scheduling problems.

'Danger Theory: The Link between AIS and IDS?'

Abstract We present ideas about creating a next generation Intrusion Detection System (IDS) based on the latest immunological theories. The central challenge with computer security is determining the difference between normal and potentially harmful activity. For half a century, developers have protected their systems by coding rules that identify and block specific events. However, the nature of current and future threats in conjunction with ever larger IT systems urgently requires the development of automated and adaptive defensive tools. A promising solution is emerging in the form of Artificial Immune Systems (AIS): The Human Immune System (HIS) can detect and defend against harmful and previously unseen invaders, so can we not build a similar Intrusion Detection System (IDS) for our computers? Presumably, those systems would then have the same beneficial properties as HIS like error tolerance, adaptation and self-monitoring. Current AIS have been successful on test systems, but the algorithms rely on self-nonself discrimination, as stipulated in classical immunology. However, immunologist are increasingly finding fault with traditional self-nonself thinking and a new 'Danger Theory' (DT) is emerging. This new theory suggests that the immune system reacts to threats based on the correlation of various (danger) signals and it provides a method of 'grounding' the immune response, i.e. linking it directly to the attacker. Little is currently understood of the precise nature and correlation of these signals and the theory is a topic of hot debate. It is the aim of this research to investigate this correlation and to translate the DT into the realms of computer security, thereby creating AIS that are no longer limited by self-nonself discrimination. It should be noted that we do not intend to defend this controversial theory per se, although as a deliverable this project will add to the body of knowledge in this area. Rather we are interested in its merits for scaling up AIS applications by overcoming self-nonself discrimination problems.