Just enough learning (of association rules): the TAR2 “Treatment” learner

An over-zealous machine learner can automatically generate large, intricate, theories which can be hard to understand. However, such intricate learning is not necessary in domains that lack complex relationships. A much simpler learner can suffice in domains with narrow funnels; i.e. where most domain variables are controlled by a very small subset. Such a learner is TAR2: a weighted-class minimal contrast-set association rule learner that utilizes confidence-based pruning, but not support-based pruning. TAR2 learns treatments; i.e. constraints that can change an agent’s environment. Treatments take two forms. Controller treatments hold the smallest number of conjunctions that most improve the current state of the system. Monitor treatments hold the smallest number of conjunctions that best detect future faulty system behavior. Such treatments tell an agent what to do (apply the controller) and what to watch for (the monitor conditions) within the current environment. Because TAR2 generates very small theories, our experience has been that users prefer its tiny treatments. The success of such a simple learner suggests that many domains lack complex relationships.     

An experimental test of Occam’s razor in classification

A widely persisting interpretation of Occam’s razor is that given two classifiers with the same training error, the simpler classifier is more likely to generalize better. Within a long-lasting debate in the machine learning community over Occam’s razor, Domingos (Data Min. Knowl. Discov. 3:409–425, 1999) rejects this interpretation and proposes that model complexity is only a confounding factor usually correlated with the number of models from which the learner selects. It is thus hypothesized that the risk of overfitting (poor generalization) follows only from the number of model tests rather than the complexity of the selected model. We test this hypothesis on 30 UCI data sets using polynomial classification models. The results confirm Domingos’ hypothesis on the 0.05 significance level and thus refutes the above interpretation of Occam’s razor. Our experiments however also illustrate that decoupling the two factors (model complexity and number of model tests) is problematic.     

Memory Consolidation and Forgetting During Sleep: A Neural Network Model

Experimental research with humans and animals suggests that sleep — particularly REM sleep — is, in some way, associated with learning. However, the nature of the association and the underlying mechanism remain unclear. A number of theoretical models have drawn inspiration from research into Artificial Neural Networks. Crick and Mitchinson's ‘unlearning’ and Robins and McCallum's ‘pseudo-rehearsal’ models suggest alternative mechanisms through which sleep could contribute to learning. In this paper we present simulations, suggesting a possible synthesis. Our simulations use a modified version of a Hopfield network to model the possible contribution of sleep to memory consolidation. Sleep is simulated by removing all sensory input to the network and by exposing it to a ‘noise’, intended as a highly abstract model of the signals generated by the Ponto-geniculate-occipital system during sleep. The results show that simulated sleep does indeed contribute to learning and that the relationship between the observed effect and the length of simulated sleep can be represented by a U-shaped curve. It is shown that while high-amplitude, low-frequency noise (reminiscent of NREM sleep) leads to a general reinforcement of memory, low-amplitude, high-frequency noise (as observed in REM sleep) leads to ‘forgetting’ of all but the strongest memory traces. This suggests that a combination of the two kinds of sleep might produce a stronger effect than either kind of sleep on its own and that effective consolidation of memory during sleep may depend not just on REM or NREM sleep but on the overall dynamics of the sleep cycle.     

Rethinking the cross-cultural interaction architecture

The paper is an exploration for a conceptual framework for cross-cultural interfacing. The roots of this exploration lie in my personal, functional, social and cultural experiences, and cross-cultural encounters. These encounters in many ways reflect the networking journey of AI & Society, promoting and stimulating the human-centred debate in cross-cultural settings. As a ‘cross-cultural holon’, AI & Society has been questioning the given orthodoxy of the ‘one best way’ and the culture of the ‘exact language’ since its inception 21 years ago. My observation is that even though we now live in a world of multimodal interactive media technologies, some how the interface design has not moved much beyond the functional (observed reality) worldview of our interactions. Even at the functional level, the design perspective remains bounded by the observed reality. It is thus no surprise that when it comes to designing interaction technologies for social and cultural domains such as health care, work, employment, life long learning, the design mainly hovers around the functional skill training, mechanical feedback and objectified evaluation. When we move beyond the functional worldview, and explore our interactions within social and cultural domains, we encounter both the reality (objectified world) and actuality (experienced world) of interactions. At this level, the challenge of the interface is not just coping with the interaction spaces of reality and actuality but also with the ‘in-between’ space of actuality–reality gaps. We explore how the concept of symbiosis enables the interaction in the gaps between actuality and reality, how the concept of the ‘culture of the artificial enables the sharing and pooling experiences, how the concept of valorisation’ enables to find a coherence (commonality) between interactions, and how the concept of cultural holon is used to conceptualise the network architecture of cross-cultural interactions. It is proposed that the way forward to designing cross-cultural interfaces is to first seek a conceptual framework for cross-cultural interaction spaces, and then use this framework to design interfacing systems and tools.     

Chasing a Definition of “Alarm”

Alarm management has been around for decades in telecom solutions. We have seen various efforts to define standardised alarm interfaces. The research community has focused on various alarms correlation strategies. Still, after years of effort in industry and research alike, network administrators are flooded with alarms; alarms are suffering from poor information quality; and the costs of alarm integration have not decreased. In this paper, we explore the concept of ‘alarm’. We define ‘alarm’ and alarm-type concepts by investigating the different definitions currently in use in standards and research efforts. Based on statistical alarm data from a mobile operator we argue that operational and capital expenditures would decrease if alarm sources would apply to our alarm model.     

What the <0.70, 1.17, 0.99, 1.07> is a Symbol?

The notion of a ‘symbol’ plays an important role in the disciplines of Philosophy, Psychology, Computer Science, and Cognitive Science. However, there is comparatively little agreement on how this notion is to be understood, either between disciplines, or even within particular disciplines. This paper does not attempt to defend some putatively ‘correct’ version of the concept of a ‘symbol.’ Rather, some terminological conventions are suggested, some constraints are proposed and a taxonomy of the kinds of issue that give rise to disagreement is articulated. The goal here is to provide something like a ‘geography’ of the various notions of ‘symbol’ that have appeared in the various literatures, so as to highlight the key issues and to permit the focusing of attention upon the important dimensions. In particular, the relationship between ‘tokens’ and ‘symbols’ is addressed. The issue of designation is discussed in some detail. The distinction between simple and complex symbols is clarified and an apparently necessary condition for a system to be potentially symbol, or token bearing, is introduced.     

Multi-user mixed reality system ‘Gulliver’s World’: a case study on collaborative edutainment at the intersection of material and virtual worlds

This case study is about ‘Gulliver’s World’, a multi-user mixed reality environment that functions simultaneously as interactive edutainment platform and learning environment as well as flexible infrastructure for the expansion of mixed reality environments via innovations in technology and media art. As an exhibition project, the installation is characterized by a nonlinear exhibition concept that activates interaction between individual users and different modes of virtual reality as well as collaboration among the users themselves. At seven workstations, people of all age groups range along the Reality–Virtuality continuum while collaboratively creating 3D worlds. Results of these creation activities are interactive worlds at the nexus of theatre, digital film production and game environment. As a research project, ‘Gulliver’s World’ features multilevel infrastructure with exemplary content in which the latest insights and models to emerge from HCI research as well as concepts of mixed reality and virtual environments and their supporting technology are brought together and developed further.     

Institution Morphisms

Institutions formalise the intuitive notion of logical system, including syntax, semantics, and the relation of satisfaction between them. Our exposition emphasises the natural way that institutions can support deduction on sentences, and inclusions of signatures, theories, etc.; it also introduces terminology to clearly distinguish several levels of generality of the institution concept. A surprising number of different notions of morphism have been suggested for forming categories with institutions as objects, and an amazing variety of names have been proposed for them. One goal of this paper is to suggest a terminology that is uniform and informative to replace the current chaotic nomenclature; another goal is to investigate the properties and interrelations of these notions in a systematic way. Following brief expositions of indexed categories, diagram categories, twisted relations and Kan extensions, we demonstrate and then exploit the duality between institution morphisms in the original sense of Goguen and Burstall, and the ‘plain maps’ of Meseguer, obtaining simple uniform proofs of completeness and cocompleteness for both resulting categories. Because of this duality, we prefer the name ‘comorphism’ over ‘plain map’; moreover, we argue that morphisms are more natural than comorphisms in many cases. We also consider ‘theoroidal’ morphisms and comorphisms, which generalise signatures to theories, based on a theoroidal institution construction, finding that the ‘maps’ of Meseguer are theoroidal comorphisms, while theoroidal morphisms are a new concept. We introduce ‘forward’ and ‘semi-natural’ morphisms, and develop some of their properties. Appendices discuss institutions for partial algebra, a variant of order sorted algebra, two versions of hidden algebra, and a generalisation of universal algebra; these illustrate various points in the main text. A final appendix makes explicit a greater generality for the institution concept, clarifies certain details and proves some results that lift institution theory to this level. Received December 2000 / Accepted in revised form January 2002     

Voting techniques for expert search

In an expert search task, the users’ need is to identify people who have relevant expertise to a topic of interest. An expert search system predicts and ranks the expertise of a set of candidate persons with respect to the users’ query. In this paper, we propose a novel approach for predicting and ranking candidate expertise with respect to a query, called the Voting Model for Expert Search. In the Voting Model, we see the problem of ranking experts as a voting problem. We model the voting problem using 12 various voting techniques, which are inspired from the data fusion field. We investigate the effectiveness of the Voting Model and the associated voting techniques across a range of document weighting models, in the context of the TREC 2005 and TREC 2006 Enterprise tracks. The evaluation results show that the voting paradigm is very effective, without using any query or collection-specific heuristics. Moreover, we show that improving the quality of the underlying document representation can significantly improve the retrieval performance of the voting techniques on an expert search task. In particular, we demonstrate that applying field-based weighting models improves the ranking of candidates. Finally, we demonstrate that the relative performance of the voting techniques for the proposed approach is stable on a given task regardless of the used weighting models, suggesting that some of the proposed voting techniques will always perform better than other voting techniques. Extended version of ‘Voting for candidates: adapting data fusion techniques for an expert search task’. C. Macdonald and I. Ounis. In Proceedings of ACM CIKM 2006, Arlington, VA. 2006. doi: 10.1145/1183614.1183671.     

Study on IAC networks models: learning and stability

This paper presents a methodology that uses evolutionary learning in training ‘A’ model networks, a topology based on Interactive Activation and Competition (IAC) neural networks. IAC networks show local knowledge and processing units clustered in pools. The connections among units may assume only 1, 0 or −1. On the other hand, ‘A’ model network uses values in interval [−1, 1]. This feature provides a wider range of applications for this network, including problems which do not show mutually exclusive concepts. However, there is no algorithm to adjust the network weights and still preserve the desired characteristics of the original network. Accordingly, we propose the use of genetic algorithms in a new methodology to obtain the correct weight set for this network. Two examples are used to illustrate the proposed method. Findings are considered consistent and generic enough to allow further applications on similar classes of problems suitable for ‘A’ model IAC Networks.     

Decision Making and Management of Dynamic Risk

Managing dynamic environments often requires decision making under uncertainty and risk. Two types of uncertainty are involved: uncertainty about the state and the evolution of the situation, and ‘openness’ of the possible actions to face possible consequences. In an experimental study on risk management in dynamic situations, two contrasted ‘ecological’ scenarios – transposed from effective situations of emergency management – were compared in order to identify the impact of their ‘openness’ in the subjects’ strategies for decision making. The ‘Lost Child’ scenario presented qualitative and irreversible consequences (child’s death) and high uncertainty; it exerted high demands both in risk assessment (risk representation) and action elaboration and choice. A less open situation (‘Hydrocarbon Fire’) required a main choice between two contrasted actions, with quantitative computable consequences. The strategies of ‘experimental subjects’ (university students) and ‘operative subjects’ (professional fire-fighter officers) were compared in order to evaluate the ecological validity of experimental research in this field, from the point of view of the subjects themselves. The two scenarios appeared to be independent, so that quite different models of decision making have to be hypothesised, differing by the importance of assessing risk and defining possible actions on the one hand, and by the process of choice on the other. ‘Experimental’ subjects dramatically differed from ‘operative’ subjects when confronted with the same scenario, particularly for the less technical but more demanding scenario. It is hypothesised that three components might account for the effect of the situations and for the differences between and within groups of subjects: importance of situation assessment, spatial abilities, and global orientation of activity in managing dynamic risk.     

Relations between Gold-style learning and query learning

Different formal learning models address different aspects of human learning. Below we compare Gold-style learning—modelling learning as a limiting process in which the learner may change its mind arbitrarily often before converging to a correct hypothesis—to learning via queries—modelling learning as a one-shot process in which the learner is required to identify the target concept with just one hypothesis. In the Gold-style model considered below, the information presented to the learner consists of positive examples for the target concept, whereas in query learning, the learner may pose a certain kind of queries about the target concept, which will be answered correctly by an oracle (called teacher). Although these two approaches seem rather unrelated at first glance, we provide characterisations of different models of Gold-style learning (learning in the limit, conservative inference, and behaviourally correct learning) in terms of query learning. Thus we describe the circumstances which are necessary to replace limit learners by equally powerful one-shot learners. Our results are valid in the general context of learning indexable classes of recursive languages. This analysis leads to an important observation, namely that there is a natural query learning type hierarchically in-between Gold-style learning in the limit and behaviourally correct learning. Astonishingly, this query learning type can then again be characterised in terms of Gold-style inference.     

Relational IBL in classical music

It is well known that many hard tasks considered in machine learning and data mining can be solved in a rather simple and robust way with an instance- and distance-based approach. In this work we present another difficult task: learning, from large numbers of complex performances by concert pianists, to play music expressively. We model the problem as a multi-level decomposition and prediction task. We show that this is a fundamentally relational learning problem and propose a new similarity measure for structured objects, which is built into a relational instance-based learning algorithm named DISTALL. Experiments with data derived from a substantial number of Mozart piano sonata recordings by a skilled concert pianist demonstrate that the approach is viable. We show that the instance-based learner operating on structured, relational data outperforms a propositional k-NN algorithm. In qualitative terms, some of the piano performances produced by DISTALL after learning from the human artist are of substantial musical quality; one even won a prize in an international ‘computer music performance’ contest. The experiments thus provide evidence of the capabilities of ILP in a highly complex domain such as music. Editors: Tamás Horváth and Akihiro Yamamoto     

e-Business trust concepts based on seals and insurance solutions

Consumers have embraced the concept of e-commerce although less enthusiastically than expected. Major concerns still exist regarding the use of the Internet for private purchasing. Trust is seen as a factor that is becoming increasingly important for both consumers and content managers alike. Various trust-related support features for online transactions are available, but most lack any form of guarantee or insurance for the parties involved. In this context, the paper seeks to explore the concepts and potential contributions of contract-based guarantees and insurance services with regard to business-to-consumer online transactions. After an inventory-taking of available seals of approval and insurance solutions for B2C online transactions, the paper drafts a research framework for investigating different insurance models. The case of ‘Trusted Shops’, backed by a German insurance provider, illustrates the concept of insurance solutions and analyzes benefits and risks for all parties involved. The potential advantages and limitations of extending the concept along the dimensions ‘scale’ and ‘scope’ are presented. The paper concludes by providing some suggestions for further research.     

Learning to trust in the competence and commitment of agents

For agents to collaborate in open multi-agent systems, each agent must trust in the other agents’ ability to complete tasks and willingness to cooperate. Agents need to decide between cooperative and opportunistic behavior based on their assessment of another agents’ trustworthiness. In particular, an agent can have two beliefs about a potential partner that tend to indicate trustworthiness: that the partner is competent and that the partner expects to engage in future interactions. This paper explores an approach that models competence as an agent’s probability of successfully performing an action, and models belief in future interactions as a discount factor. We evaluate the underlying decision framework’s performance given accurate knowledge of the model’s parameters in an evolutionary game setting. We then introduce a game-theoretic framework in which an agent can learn a model of another agent online, using the Harsanyi transformation. The learning agents evaluate a set of competing hypotheses about another agent during the simulated play of an indefinitely repeated game. The Harsanyi strategy is shown to demonstrate robust and successful online play against a variety of static, classic, and learning strategies in a variable-payoff Iterated Prisoner’s Dilemma setting.     

Learning Organisations: The Process of Innovation and Technological Change

In the present scenario of globalisation, knowledge has become the prime factor of production for competitive advantage. This calls for acquisition and utilisation of knowledge for innovation and technical change on a constant basis, which is only possible in a ‘learning organisation’. Innovative activities of a learning organisation are influenced by three main factors: (1) internal learning; (2) external learning; and (3) the innovation strategies decided upon by the enterprise management. An assumption has been made that, particularly in developing countries, absorption and adaptation of technologies, i.e. indigenisation, take place through a process of ‘learning by doing’. Taking this into consideration, this paper focuses on a few case studies carried out at NISTADS, New Delhi, India, on small enterprises in the formal as well as traditional sectors, highlighting the learning process in an organisational context and how it brings in innovation and technological change at enterprise level. The study demonstrates that the learning environment in an organisational context is an indispensable process to be innovative and building up capabilities for technological change. This in turn also calls for strong networking of the enterprises with academia, R&D institutions and other enterprises, to create knowledge clusters. This builds up a strong case for a network approach of learning organisations not only at the regional level but also at the cross-cultural level for constant innovation and technical change.     

Machine intelligence (MI), competence and creativity

In mid-twentieth century, the hypothesis, ‘a machine can think’ became very popular after, Alan Turing’s article on ‘Computing Machinery and Intelligence’. This hypothesis, ‘a machine can think’ established the foundations of machine intelligence (MI), and claimed that machines have consciousness and creativity, with the power to compete with human beings. In the first section, I shall show how consciousness and creativity is conceptualized in the domain of MI. The main aim of MI is not only to construct difficult programs to solve our day-to-day problems, but also to reproduce mentality in machines. It was claimed that all the mental qualities are ascribable to machines. My intention in this paper is to clarify consciousness and creativity from a subjective point of view. My claim is that consciousness and creativity cannot be fully represented in a mechanistic domain. There are subjective mental or conscious states, which can be seen in a first-person perspective of their proper understanding.     

Some Kind of Analogtivity: anti-simulation through design

This paper discusses and explores issues surrounding current approaches to the design of technological products and offers two critical design proposals for presentation and debate. Primarily driven by contemporary theoretical writings and thoughts on the subject of ‘technology’ and ‘simulation’, currently being offered by leading thinkers on these subjects and expressed as ‘technological objects’; they are the result of a critical investigation into the emerging design issues surrounding ‘interaction’ and ‘transparency’. By using ‘popular’ language of product design as a vehicle, they exist as ‘cultural offerings’ exploring an alternative future for technological products not necessarily governed by science and economics.     

Learning goal hierarchies from structured observations and expert annotations

We describe a relational learning by observation framework that automatically creates cognitive agent programs that model expert task performance in complex dynamic domains. Our framework uses observed behavior and goal annotations of an expert as the primary input, interprets them in the context of background knowledge, and returns an agent program that behaves similar to the expert. We map the problem of creating an agent program on to multiple learning problems that can be represented in a “supervised concept learning’’ setting. The acquired procedural knowledge is partitioned into a hierarchy of goals and represented with first order rules. Using an inductive logic programming (ILP) learning component allows our framework to naturally combine structured behavior observations, parametric and hierarchical goal annotations, and complex background knowledge. To deal with the large domains we consider, we have developed an efficient mechanism for storing and retrieving structured behavior data. We have tested our approach using artificially created examples and behavior observation traces generated by AI agents. We evaluate the learned rules by comparing them to hand-coded rules. Editor: Rui Camacho