A Little AI Goes a Long Way on Wall Street

This is a history and technical overview of one of the major AI successes in securities trading. It incorporates only as much AI as the financial user community could deal with, integrating tightly with their electronic environment. Computational resources are used to make a simple, highly application-specific user interface, a small firm has succeeded in developing a system used by many of the largest institutional investors and money managers in the USA. The system is directly linked to the NYSE and other electronic equity-execution systems. This is not a prototype or a proposal. This is real and in wide use today, often generating transaction volumes exceeding 5 million shares per day. Traders use a special-purpose rule-based language to describe a wide variety of market conditions. The system keeps up with high-speed incoming market data in real-time and tells the traders when, and how strongly, their specified conditions match the current state of the market. It can also formulate trading recommendations based on those conditions. Finally, and most importantly, it provides direct access to electronic execution channels for quick action on these recommendations. The combination of tight integration with both market data and electronic executions, an appropriate, accessible user interface, and a high level of support have contributed to a major AI success story. The transactions flowing through these systems produce more revenue on a good day than many other AI applications will generate over their entire operational lifetimes.     

APX: An Integrated Knowledge-Based System to Support Audit Planning

Auditors who must express an opinion on the financial statements of publicly held enterprises must perform a series of risk assessments as the basis for selecting appropriate audit procedures to be performed. Although many auditing expert systems have been developed for individual components of the audit risk assessment process such as inherent and control risk assessment, none integrate these assessments to arrive at the ultimate objective of assessing the allowable level of detection risk which drives audit procedure selection. This paper discusses critical design concerns for audit planning expert systems in light of the integrated nature of the steps involved in the audit planning process. The design and development of an integrated audit planning expert system called APX (Audit Planning eXpert) is also discussed. A distinguishing feature of APX is that it not only performs the individual types of risk assessments but it also integrates the resulting assessments to arrive at allowable detection risk. The knowledge acquisition and system development process are described to illustrate some of the mechanics of developing such a system.     

Case-based Reasoning and Multiple-agent Systems for Accounting Regulation Systems with Extensions

One of the areas of judgment research in accounting and financial applications is that of accounting regulation. Previously, artificial intelligence efforts at modeling human judgment in accounting regulation systems have concentrated on rule-based expert systems. In those systems, general heuristic knowledge was captured using ‘if … then …’ rules in order to model particular decision processes. Recent developments in artificial intelligence have focused on case-based reasoning (CBR) and multiple-agent intelligent systems (MAIS). The ideas behind CBR are that ‘if it worked once then remember to use it again’ and ‘if it did not work before, then remember to not use it again‘. MAIS assumes that many organizational systems can be treated as computational models of multiple-interacting intelligent agents. Typically, solutions may be derived using some form of negotiations between agents to accomplish single global or separate individual interacting goals. This paper argues that many accounting regulation judgment processes can be modeled using CBR and MAIS. As a result, it summarizes some examples of both CBR and MAIS useful in accounting regulation and extends those to other accounting applications. In addition, it describes the results of some previously developed systems that employ CBR or MAIS.     

Automated Induction of Rule-based Neural Networks from Databases

This paper describes our approach to the problem of automated knowledge acquisition from large databases of examples using an information-theoretic approach. Our previous research has resulted in practical algorithms (ITRULE) for the automatic induction of rules from large example databases. Utilizing these algorithms, the raw data can be transformed into a set of human readable IF THEN rules, thus giving insight into the knowledge hidden within the data. These rules can then be automatically loaded into an expert system shell. Alternatively, they can be used to build a new type of parallel inference system—a rule-based neural network. This process enables a prototype expert system to be automatically generated and up and running in a matter of minutes, compared with months using a manual knowledge-acquisition approach. The resulting expert system can then be used as a sophisticated search and analysis tool to query the original database capable of reasoning with uncertain and incomplete data.     

Anticipating transitions beyond the current mobility regimes: How acceptability matters

Today's car-based transportation systems require a transition toward sustainability. This is particularly the case in suburban areas, where the costs for introducing a new transportation system are high due to the low population density. At the same time the negative externalities of the current mobility regime – such as health costs and congestions – are increasing rapidly. Based on expert interviews with car manufacturers, transportation authorities, environmental groups, and scientists we identify two visionary characteristics of future, more sustainable transportation systems: automated driving and sharing. Using these two characteristics, we apply the scenario-axes technique to develop four mobility scenarios for a suburban context that range from business-as-usual to a radical and more sustainable one. In an evaluation with ten criteria that measure a scenario's performance from a user perspective, the radical scenario performs worst since it does not meet current individualistic user requirements. Our findings suggest that lock-ins of users’ expectations act as barriers for the diffusion of novel transportation systems. These barriers cannot be overcome by technological innovations and regulation alone. Hence, we call for innovative arenas, wherein technology and user acceptability could co-evolve.     

Refining the behavior of multiple expert systems: a concept and empirical results

Expert System Refinement (ESR) is introduced as a means to automatically refine the performance of one or more expert systems. The ESR concept is based on Holland’s learning classifier systems and a method for integrating multiple expert systems. Through users’ feedback about the usefulness/correctness of the integrated expert system’s recommendations, ESR enables behaviors of both individual expert systems as well as the integrated system to improve over time. The ESR concept is tested on a German Credit Database. This empirical evidence suggests that the ESR concept can be usefully applied in automating the process of expert system refinement and multiple expert systems integration. © 1998 John Wiley & Sons, Ltd.     

An Analysis of Potential Legal Liability Incurred Through Audit Expert Systems

The introduction of expert systems technology into the audit environment has opened a new avenue of auditor legal liability. This paper examines the potential impact expert systems will have on auditor liability. The presentation of this new avenue of auditors' legal liability explores both the potential for litigation under failure of auditor/expert system collaboration to yield prudent decisions and the failure to use an available expert system. The risks evolving from failure to use an available expert system include the possibility that the system could be used against the auditor in the courtroom. While case law will ultimately determine the bounds of this liability, this paper acquaints the reader with the important legal issues involved and the varied outcomes that could emerge. It should also be noted that while the specific example presented in this paper relates to the audit profession, the legal concepts are of equivalent concern to other professions enduring broad implementation of expert systems.     

Combining Expert Systems and Hypertext: Opportunities and Obstacles

Hypertext discussions are occurring more frequently at expert systems conferences. Hypertext is not an expert system language or expert system shell, but instead it is simultaneously a method of storing and retrieving data. The growing interest in hypertext in the expert system domain is because the combination of hypertext capabilities and expert systems creates a compelling synergistic relationship (Oren, 1987). During a panel discussion at the Second International Symposium on Expert Systems in Business, Finance and Accounting, Bill Swartout compared traditional expert systems to hypertext by saying that the knowledge base (e.g. the production rules) of an expert system can be thought of as formal knowledge and hypertext can be thought of as informal knowledge. This article will demonstrate how this informal knowledge can add power to traditional expert systems by increasing the flexibility of the explanation facilities and thereby the appeal of the system to a broader range of users. The article also will discuss how expert systems can improve the functionality of hypertext systems by adding intelligence to traditional hypertext systems. This article is divided into three sections. The first describes basic hypertext concepts. The second section shows how hypertext can enhance expert systems and, conversely, how expert systems can enhance hypertext systems. The final section discusses some potential problems and concerns that must be considered when designing a hypertext system.     

Expert Systems at the Customer Interface

ESCFE (Expert System for Customer Facing Environments) is an integrated expert systems architecture which supports a three-way dialogue between British Telecommunications (BT) customers, staff and its technical and commercial support systems. ESCFE is customer facing—this means that the user interface dialogue is directed explicitly at the customer, as opposed to other expert systems/help desks which have a user interface directed at staff, who then deal with the customer in whatever manner they have been trained. In dealing with the customer, BT staff require a wide range of knowledge concerning the products and services portfolio, internal BT procedures, selling opportunities, customer-care policy and contractual liabilities. This knowledge can be contained within ESCFE. The knowledge is used to support a dialogue which allows for both customer- and staff-driven interaction, and for different levels of expertise in both customers and staff.     

ICARE: A Knowledge-Based Underwriting System

This paper reports on the development of ICARE (Intelligent Computer Assisted Risk Evaluation), a knowledge-based system for underwriting personal lines of insurance at a large insurance company. ICARE uses rule-based and object-oriented programming techniques to implement the subjective decision processes of expert underwriters for automobile insurance lines. The system is an easy-to-understand, easy-to-maintain, and efficient software program in which nontechnical underwriters can maintain their own knowledge base of underwriting rules and guidelines.     

Expert Systems for Evaluating Business Opportunities: Implementing the Management Advisor at Krypton Chemical

This paper investigates issues arising in procuring and implementing a commercial high-end expert system, Palladian's Management Advisor. This system is intended for improving corporate decisions about financial investments in new business opportunities, e.g. projects involving new plants, equipment, products, or marketing expenditures. We look at Krypton Chemical's experiences with the Management Advisor from both the customer's and the vendor's viewpoints. The discussion raises some difficult questions about what it takes to make high-end financial expert systems successful. The paper then concludes with recommendations intended to improve design, marketing, procurement and implementation practices with respect to such systems.     

Machine‐independent audit trail analysis—a tool for continuous audit assurance

This paper reports the results of a research project which examines the feasibility of developing a machine‐independent audit trail analyser (MIATA). MIATA is a knowledge‐based system which performs intelligent analysis of operating system audit trails. Such a system is proposed as a decision support tool for auditors when assessing the risk of unauthorized user activity in multi‐user computer systems. It is also relevant to the provision of a continuous assurance service to clients by internal and external auditors. Monitoring user activity in system audit trails manually is impractical because of the vast quantity of events recorded in those audit trails. However, if done manually, an expert security auditor would be needed to look for two main types of events—user activity rejected by the system's security settings (failed actions) and users behaving abnormally (e.g. unexpected changes in activity such as the purchasing clerk attempting to modify payroll data). A knowledge‐based system is suited to applications that require expertise to perform well‐de?ned, yet complex, monitoring activities (e.g. controlling nuclear reactors and detecting intrusions in computer systems). To permit machine‐independent intelligent audit trail analysis, an anomaly‐detection approach is adopted. Time series forecasting methods are used to develop and maintain the user pro?le database (knowledge base) that allows identi?cation of users with rejected behaviour as well as abnormal behaviour. The knowledge‐based system maintains this knowledge base and permits reporting on the potential intruder threats (summarized in Table I). The intelligence of the MIATA system is its ability to handle audit trails from any system, its knowledge base capturing rejected user activity and detecting anomalous activity, and its reporting capabilities focusing on known methods of intrusion. MIATA also updates user pro?les and forecasts of behaviour on a daily basis. As such, it also ‘learns’ from changes in user behaviour. The feasibility of generating machine‐independent audit trail records, and the applicability of the anomaly‐detection approach and time series forecasting methods, are demonstrated using three case studies. These results support the proposal that developing a machine‐independent audit trail analyser is feasible. Such a system will be an invaluable aid to an auditor in detecting potential computer intrusions and monitoring user activity. Copyright © 2004 John Wiley & Sons, Ltd.     

Future utility services’ (un)knowns framework: Knowledge existence and knowledge reach

This paper presents a framework for assessing risks and uncertainties in the domain of utility services in futures up to 100 years hence. A nine-step process framework is presented following a review of extant literature. The framework provides a way to quantitatively and qualitatively evaluate one or more vignettes (potential future utility infrastructure solutions) by examining: 1. Alternative scenario contexts; 2. User utility service requirements; 3. Technologies’ readiness; 4. Knowledge of each technology life cycle stage; 5. Criticality of technologies to user requirements; 6. Interdependencies between technologies. The framework is based on collecting multiple expert contributions in order to arrive at a comparative evaluation. We use the ‘city blood’ vignette, which represents an infrastructure system delivering hydrogen enriched water, to apply the framework to two radically different scenario contexts.     

Information technology and the board of directors

Ever since the Y2K scare, boards have grown increasingly nervous about corporate dependence on information technology. Since then, computer crashes, denial of service attacks, competitive pressures, and the need to automate compliance with government regulations have heightened board sensitivity to IT risk. Unfortunately, most boards remain largely in the dark when it comes to IT spending and strategy, despite the fact that corporate information assets can account for more than 50% of capital spending. A lack of board oversight for IT activities is dangerous, the authors say. It puts firms at risk in the same way that failing to audit their books would. Companies that have established board-level IT governance committees are better able to control IT project costs and carve out competitive advantage. But there is no one-size-fits-all model for board supervision of a company's IT operations. The correct approach depends on what strategic "mode" a company is in whether its operations are extremely dependent on IT or not, and whether or not it relies heavily on keeping up with the latest technologies. This article spells out the conditions under which boards need to change their level of involvement in IT decisions, explaining how members can recognize their firms' IT risks and decide whether they should pursue more aggressive IT governance. The authors delineate what an IT governance committee should look like in terms of charter, membership, duties, and overall agenda. They also offer recommendations for developing IT policies that take into account an organization's operational and strategic needs and suggest what to do when those needs change. Given the dizzying pace of change in the world of IT, boards can't afford to ignore the state of their IT systems and capabilities. Appropriate board governance can go a long way toward helping a company avoid unnecessary risk and improve its competitive position.     

A fuzzy logic-driven multiple knowledge integration framework for improving the performance of expert systems

To maintain a high performance in an ill-structured situation, expert systems should depend on multiple sources of knowledge rather than a single type. For this reason, we propose multiple knowledge integration by using a fuzzy logic-driven framework. Types of knowledge being considered here are threefold: machine, expert and user. Machine knowledge is obtained by a back- propagation neural network model from historical instances of a target problem domain. Expert knowledge is related to interpreting the trends of external factors that seem to affect the target problem domain. User knowledge represents a user’s personal views about information given by both expert knowledge and machine knowledge. The target problem domain of this paper is one-week-ahead stock market stage prediction: Bull, Edged-up, Edged-down, and Bear. Extensive experiments with real data proved that the proposed fuzzy logic-driven framework for multiple knowledge integration can contribute significantly to improving the performance of expert systems. Copyright © 1998 John Wiley & Sons, Ltd.     

A Hybrid-Based Expert System for Personal Pension Planning in the UK

While advanced computing technology, and particularly the use of artificial intelligence in the form of expert systems, could not necessarily be said to be common in the US financial planning domain, it is certainly not unheard of. This situation is significantly different from that found in the comparable UK domain. This paper is based on a project to look at the use of computing technology to support the role of the personal financial adviser in the UK—a domain in which little published research work has been undertaken. It briefly describes the current UK marketplace and details the construction of a small hybrid-based expert system, built to support the selection of personal pension plans, to illustrate the inherent value in developing such technology in this domain. The paper discusses the benefits of using hybrid representational techniques as opposed to single representations in creating an expert system in a financial planning domain.     

Using Constraint Programming to Design an Option-based Decision Support System

Financial options and futures give investors the opportunity to form complex strategies that meet their investment objectives for risk management. However, this opportunity gives rise to a difficult task: finding a desired strategy from among a large number of possible strategies. This paper describes an intelligent decision-support system for generating option-based investment strategies by using constraint programming, which is an integrated framework of Artificial Intelligence and Logic Programming. In this system, constraint programming acts as a bridge between qualitative and quantitative analyses in decision processes. In qualitative analysis, logical reasoning with hypotheses is used to automatically create complex strategies through abstract matching with investors' profiles. Here, abstract matching can be regarded as symbolic computation for producing qualitatively reasonable strategies. In quantitative analysis, a set of complete solutions that satisfy user-supplied constraints are obtained by constraint satisfaction and optimization. A constraint language based on the framework of Constraint Logic Programming has been developed in order to integrate these symbolic and numerical computations in a uniform way. The resulting system written in this language has the following features. (1) Unlike rule-based expert systems, the constraint-based system can create novel investment strategies. (2) A smooth transition from qualitative to quantitative analyses can be naturally achieved due to the constraint language. (3) Qualitative analysis can reduce search complexity, because the analysis focuses on a small set of qualitative distinctions in solution space. These features indicate the usefulness of constraint programming for designing intelligent decision-support systems.     

Decision‐making capabilities of a hybrid system applied to the auditor's going‐concern assessment

The purpose of this study is to evaluate a hybrid system as a decision support model to assist with the auditor's going‐concern assessment. The going‐concern assessment is often an unstructured decision that involves the use of both qualitative and quantitative information. An expert system that predicts the going‐concern decision has been developed in consultation with partners at three of the Big Five accounting firms. This system is combined with a statistical model that predicts bankruptcy, as a component of the auditor's decision, to form a hybrid system. The hybrid system, because it combines the use of quantitative and qualitative information, has the potential for better prediction accuracy than either the expert system or statistical model predicting separately. In addition, testing of the system provides some insight into the characteristics of firms that experience problems, but do not necessarily receive a going‐concern modification. Further investigation into those firms that have problems could reveal factors that may be incorporated into decision support systems for auditors, in order to improve accuracy and reliability of these decision tools. © 2001 John Wiley & Sons, Ltd.     

SCAN: A Case-Based Reasoning Model for Generating Information System Control Recommendations

This paper describes SCAN, a case-based reasoning model for generating information system control recommendations. The purpose of the paper is to explain how a case-based reasoner may be used to support inexperienced information system auditors in evaluating controls and proposing audit recommendations. As a case-based reasoner, SCAN functions by reasoning by analogy to similar past cases. SCAN models audit experience as traces of past cases which are stored in a case knowledge base. In addition to a database of past audit cases, SCAN consists of indices for storing and retrieving cases, a similarity metric for measuring case similarity, and rules for using similar cases to generate control recommendations. SCAN uses past cases to remind the user of previous control failures, to set expectations about case features and controls, to use as a pattern against which to compare a client's controls and to help justify or explain its recommendations. SCAN's recommendations were judged to be more like those of an experienced auditor than either a student or a textbook model.     

Expert judgement and lethal toxicity of inhaled chemicals

The application of formal expert judgement for assessing quantitative data of the toxicity of large amounts of inhaled chemical substances after a potential major hazards incident is described. In this particular application, the expert assessments were used to derive probit relations for the acute lethal effects of five hazardous substances: ammonia, acrylonitrile, hydrogen fluoride, sulphur trioxide and azinphos-methyl. The purpose of the study was to develop a protocol for selecting experts and the elicitation and analysis of expert assessments, to apply this protocol to generate probit relations for representative substances, and to evaluate the overall performance of the protocol. The use of formal expert judgement includes quantitative estimates of variables, calibration of experts and representation of uncertainty on the variables. This enables analysts to optimize combined experts' assessments for establishing quantitative results for further use in probit relations. Twenty-seven experts, distributed over the five substances, gave medians and 90% central confidence bands of the lethal dose-response relations under several conditions. The calculated dose-response relations are shown and compared to existing probit relations with respect to its impact on risk assessment outputs. The calculated relations can be applied in quantitative risk assessments, for example, under the Dutch law on major hazards installations. The scientific and practical efforts and costs of using the expert judgement technique are described.