Robert Elliott Smith

Robert Elliott Smith
University College London | UCL · Department of Computer Science

Doctor of Philosophy

About

117
Publications
16,959
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
3,356
Citations
Additional affiliations
July 2005 - April 2016
University College London
Position
  • Senior Research Fellow and College Teacher
June 1997 - July 2005
University of the West of England, Bristol
Position
  • Managing Director
January 1991 - June 1997
University of Alabama
Position
  • Professor (Associate)

Publications

Publications (117)
Preprint
Full-text available
There is growing interest in the role of sentiment in economic decision-making. However, most research on the subject has focused on positive and negative valence. Conviction Narrative Theory (CNT) places Approach and Avoidance sentiment (that which drives action) at the heart of real-world decision-making, and argues that it better captures emotio...
Preprint
Full-text available
Models based on the transformer architecture, such as BERT, have marked a crucial step forward in the field of Natural Language Processing. Importantly, they allow the creation of word embeddings that capture important semantic information about words in context. However, as single entities, these embeddings are difficult to interpret and the model...
Article
Full-text available
Abstract In Communication Theory, intermedia agenda-setting refers to the influence that different news sources may have on each other, and how this subsequently affects the breadth of information that is presented to the public. Several studies have attempted to quantify the impact of intermedia agenda-setting in specific countries or contexts, bu...
Article
Full-text available
Online social networks provide users with unprecedented opportunities to engage with diverse opinions. At the same time, they enable confirmation bias on large scales by empowering individuals to self-select narratives they want to be exposed to. A precise understanding of such tradeoffs is still largely missing. We introduce a social learning mode...
Preprint
In Communication Theory, intermedia agenda-setting refers to the influence that different news sources may have on each other, and how this subsequently affects the breadth of information that is presented to the public. Several studies have attempted to quantify the impact of intermedia agenda-setting in specific countries or contexts, but a large...
Book
We live in a world increasingly ruled by technology; we seem as governed by technology as we do by laws and regulations. Frighteningly often, the influence of technology in and on our lives goes completely unchallenged by citizens and governments. We comfort ourselves with the soothing refrain that technology has no morals and can display no prejud...
Preprint
Full-text available
Online social media provide users with unprecedented opportunities to engage with diverse opinions. Simultaneously, they allow the spread of misinformation by empowering individuals to self-select the narratives they want to be exposed to, both through active (confirmation bias) and passive (personalized news algorithms) self-reinforcing mechanisms...
Conference Paper
We present the results of a preliminary study to test the hypothesis that it is possible to automatically identify opinions, in the form of conviction narratives, as they emerge in text data, and to measure and monitor how actors in the online news media influence others in the media to adopt similar narratives to their own. Narratives are represen...
Article
Full-text available
This paper applies algorithmic analysis to financial market text-based data to assess how narratives and sentiment might drive financial system developments. We find changes in emotional content in narratives are highly correlated across data sources and show the formation (and subsequent collapse) of exuberance prior to the global financial crisis...
Article
Full-text available
At a time when economics is giving intense scrutiny to the likely impact of artificial intelligence (AI) on the global economy, this paper suggests the two disciplines face a common problem when it comes to uncertainty. It is argued that, despite the enormous achievements of AI systems, it would be a serious mistake to suppose that such systems, un...
Article
Full-text available
The financial crisis of 2008 was unforeseen partly because the academic theories that underpin policy making do not sufficiently account for uncertainty and complexity or learned and evolved human capabilities for managing them. Mainstream theories of decision making tend to be strongly normative and based on wishfully unrealistic “idealized” model...
Conference Paper
Full-text available
2 The largely unexpected arrival of the global economic crisis and the largely unpredicted slowness of the recovery from the Great Recession should be precipitating an intellectual crisis across economics and policy making. We require additional theories and additional methods to detect how an economy is evolving and to provide the basis for policy...
Article
Full-text available
A number of recent contributions have tried to add to the understanding and forecasting of the macro economy by analysing news and narratives. In this contribution we report on a new approach to the content analysis of very large text databases. It draws on a new social-psychological theory of decision-making under uncertainty to focus content anal...
Article
Recursive Bayesian estimation using sequential Monte Carlos methods is a powerful numerical technique to understand latent dynamics of nonlinear non-Gaussian dynamical systems. It enables us to reason under uncertainty and addresses shortcomings underlying deterministic systems and control theories which do not provide sufficient means of performin...
Article
We develop social network and “relative sentiment shift” analysis techniques to study how financial narratives influence financial markets. First, we analyze Reuters News articles focusing on narratives about Fannie Mae. Second, we analyze Broadband and Energy narratives in the Enron Corporation email database. Combining datasets we show that phant...
Article
Full-text available
Social media analytics is showing promise for the prediction of financial markets. However, the true value of such data for trading is unclear due to a lack of consensus on which instruments can be predicted and how. Current approaches are based on the evaluation of message volumes and are typically assessed via retrospective (ex-post facto) evalua...
Data
Supplementary Information for When Can Social Media Lead Financial Markets?
Article
We develop social network and "relative sentiment shift" analysis techniques to study how financial narratives influence financial markets. First, we analyze Reuters News articles focusing on narratives about Fannie Mae. Second, we analyze Broadband and Energy narratives in the Enron Corporation email database. Combining datasets we show that phant...
Conference Paper
Multi-objective optimization problems consist of numerous, often conflicting, criteria for which any solution existing on the Pareto front of criterion trade-offs is considered optimal. In this paper we present a general-purpose algorithm designed for solving multi-objective problems (MOPS) on graphics processing units (GPUs). Specifically, a purel...
Conference Paper
Much research has been conducted in recent years applying support vector machines (SVMs) for financial forecasting. Financial time series have been shown to be very noisy and difficult to generalize: directional accuracies are often close to the class distribution. In order to improve the accuracy of our predictions, we look at applying rejection t...
Conference Paper
Stochastic volatility estimation is an important task for correctly pricing derivatives in mathematical finance. Such derivatives are used by varying types of market participant as either hedging tools or for bespoke market exposure. We evaluate our adaptive path particle filter, a recombinatory evolutionary algorithm based on the generation gap co...
Article
In this paper we propose hybrid metaheuristic particle filters for the dual estimation of state and parameters in a stochastic volatility estimation problem. We use evolutionary strategies and real coded genetic algorithms as the metaheuristics. The hybrid metaheuristic particle filters provide accurate results while using lesser number of particle...
Article
A novel self-governing system, which is theoretically founded on information theory, is introduced with the ability of determining the optimal quantity and connectivity of the hidden-layer of a three layer feed-forward neural network. The system - called MINES - simultaneously links parameter learning (performed by back-propagation) to structural l...
Article
Full-text available
Particle filters are an important class of online posterior density estimation algorithms. In this paper we propose a real coded genetic algorithm particle filter (RGAPF) for the dual estimation of stochastic volatility and parameters of a Heston type stochastic volatility model. We compare the performance of our hybrid particle filter with a param...
Conference Paper
Traditional sequential Monte-Carlo methods suffer from weight degeneracy which is where the number of distinct particles collapse. This is a particularly debilitating problem in many practical applications. A new method, the adaptive path particle filter, based on the generation gap concept from evolutionary computation, is proposed for recursive B...
Conference Paper
This article presents a new approach for automatically determining the optimal quantity and connectivity of the hidden-layer of a three-layer Feed-Forward Neural Network (FFNN) based on a theoretical and practical approach. The system (MINES) is a combination of Neural Network (NN), Back-Propagation (BP), Genetic Algorithm (GA), Mutual Information...
Article
Full-text available
Evolutionary based data mining techniques are increasingly applied to problems in the bioinformatics domain. We investigate an important aspect of predicting the folded D structure of proteins from their unfolded residue sequence using evolutionary based machine learning techniques. Our approach is to predict specific features of residues in folded...
Article
Full-text available
We investigate automated and generic alphabet reduction techniques for protein structure prediction datasets. Reducing alphabet cardinality without losing key biochemical information opens the door to potentially faster machine learning, data mining and optimization applications in structural bioinformatics. Furthermore, reduced but informative alp...
Data
Full-text available
Reduction groups obtained for all the training sets. This document lists 6 tables (3 reduction strategies and two datasets) containing the details of the reduction groups generated by our protocol for each of the ten training sets.
Conference Paper
This paper studies the performance of a newly developed supervised Michigan-style learning classifier system (LCS), called MILCS, on protein structure prediction problems and our observation of its default hierarchies (DHs). We present experimental results, and contrast them to results from other machine learning systems, named XCS, UCS, GAssist, B...
Article
This paper introduces a new variety of learning classifier system (LCS), called MILCS, which utilizes mutual information as fitness feedback. Unlike most LCSs, MILCS is specifically designed for supervised learning. We present experimental results, and contrast them to results from XCS, UCS, GAssist, BioHEL, C4.5 and Naïve Bayes. We discuss the exp...
Chapter
This chapter explains how structural learning performed by multi-variate estimation of distribution algorithms (EDAs) while building their probabilistic models is a form of linkage learning. We then show how multi-variate EDAs linkage learning mechanisms can be misled with the help of two test problems; the concatenated parity function (CPF), and t...
Chapter
Full-text available
We present our most recent efforts in applying XCS to automatic target recognition (ATR). We place particular emphasis on ATR as a series of linked problems, which include pre-processing of multi-spectral data, detection of objects (in this case, vehicles) in that data, and identification (classification) of those objects. Multi-spectral data conta...
Chapter
This paper introduces a new variety of learning classifier system (LCS), called MILCS, which utilizes mutual information as fitness feedback. Unlike most LCSs, MILCS is specifically designed for supervised learning. We present preliminary results, and contrast them to results from XCS. We discuss the explanatory power of the resulting rule sets and...
Chapter
In the multi-agent system (MAS) context, the theories and practices of evolutionary computation (EC) have new implications, particularly with regard to engineering and shaping system behaviors. Thus, it is important that we consider the embodiment of EC in “real” agents, that is, agents that involve the real restrictions of time and space within MA...
Conference Paper
Full-text available
This paper investigates the performance of estimation of distribution algorithms (EDAs) over binary test problems containing parity functions. We describe two test problems; the concatenated parity function (CPF), and the concatenated parity/trap function (CP/TF). Although these functions are separable, with bounded complexity and uniformly scaled...
Conference Paper
This paper introduces a new variety of learning classifier system (LCS), called MILCS, which utilizes mutual information as fitness feedback. Unlike most LCSs, MILCS is specifically designed for supervised learning. We present preliminary results, and contrast them to results from XCS. We discuss the explanatory power of the resulting rule sets, an...
Article
This work developed and demonstrated a machine learning approach for robust ATR. The primary innovation of this work was the development of an automated way of developing inference rules that can draw on multiple models and multiple feature types to make robust ATR decisions. The key realization is that this “meta learning” problem is one of struct...
Conference Paper
Full-text available
3.CLASSIFCATION RESULTS We present new results from our most recent efforts in applying XCS to automatic target recognition (ATR). We place particular emphasis on ATR as a series of linked problems, which include pre-processing of multi-spectral data, detection of objects (in this case, vehicles) in that data, and identification (classification) of...
Conference Paper
Full-text available
This paper introduces a new variety of learning classifier system (LCS), called MILCS, which utilizes mutual information as fitness feedback. Unlike most LCSs, MILCS is specifically designed for supervised learning. MILCS's design draws on an analogy to the structural learning approach of cascade correlation networks. We present preliminary results...
Conference Paper
Full-text available
We describe a k-bounded and additively separable test problem on which the hierarchical Bayesian Optimization Algorithm (hBOA) scales exponentially.
Chapter
In the multi-agent system (MAS) context, the theories and practices of evolutionary computation (EC) have new implications, particularly with regard to engineering and shaping system behaviors. Thus, it is important that we consider the embodiment of EC in “real” agents, that is, agents that involve the real restrictions of time and space within MA...
Article
This paper outlines our long-term vision for integrating robust machine learning as an approach to the modern battlefield. We will develop the architecture for an Integrated Learning System (ILS) that will enable representation tools to maximize the utility of data collected by distributed sensors. This project will suggest a system for data captur...
Article
Full-text available
The journeys that could be brought under Non-classical Computation are imposing challenges for computer sciences. Wittgenstein's works on the philosophy of languages, Tractatus, and its relationship to the world can be used as a model of classical computation. The Tractatus proposes that a proposition can have only one complete analysis dependent o...
Conference Paper
This paper introduces a new algorithm for determining the appropriate linkage between variables for an evolutionary algorithm. It operates in an iterative mode, as a pre-processing step before the evolutionary algorithm is run. The technique works by estimating the mutual information between variables, based on truncation-selected groups from rando...
Article
Full-text available
We propose that bio-inspired algorithms are best developed and analysed in the context of a multidisciplinary conceptual framework that provides for sophisticated biological models and well-founded analytical principles, and we outline such a framework here, in the context of Artificial Immune System (AIS) network models, and we discuss mathematica...
Conference Paper
Full-text available
A primary strength of the XCS approach is its ability to create maximally accurate general rules. In automatic target recognition (ATR) there is a need for robust performance beyond so-called standard operating conditions (SOCs, those conditions for which training data is available) to extended operating conditions (EOCs, conditions of known target...
Article
We propose that bio-inspired algorithms are best developed and analysed in the context of a multidisciplinary conceptual framework that provides for sophisticated biological models and well-founded analytical principles, and we outline such a framework here, in the context of Artificial Immune System (AIS) network models, and we discuss mathematica...
Conference Paper
Full-text available
We propose that bio-inspired algorithms are best developed and analysed in the context of a multidisciplinary conceptual framework that provides for sophisticated biological models and well-founded analytical principles, and we outline such a framework here, in the context of AIS network models. We further propose ways to unify several domains into...
Chapter
This chapter reports the authors’ ongoing experience with a system for discovering novel fighter combat maneuvers, using a genetics-based machine learning process, and combat simulation. Despite the difficulties often experienced with LCSs, this complex, real-world application has proved very successful. In effect, the adaptive system is taking the...
Conference Paper
Full-text available
We present a methodology to detect changes in quality of information (QoI) of data received by an autonomous entity. QoI is defined as the inverse of the expected Kullback-Leibler distance between a reference probability distribution and the conditional distribution associated with the data. When the underlying dynamic process that generates the da...
Conference Paper
Full-text available
The UK Grand Challenges for Computing Research is an initiative to map out certain key areas that could be used to help drive research over the next 10-15 years. One of the identified Grand Challenges is Non-Classical Computation, which examines many of the fundamental assumptions of Computer Science, and asks what would result if they were systema...
Conference Paper
Full-text available
This paper presents results and observations from the authors' con- tinuing explorations of EC systems where population members act as autono- mous agents that conduct their own, independent evaluation of (and reproduc- tion with) other agents. In particular, we consider diversity preservation in one such agent-based EC system, applied to the multi...
Conference Paper
Full-text available
While genetically inspired approaches to multi-objective optimization have many advantages over conventional approaches, they do not explicitly ex- ploit directional/gradient information. This paper describes how steepest- descent, multi-objective optimization theory can be combined with EC concepts to produce improved algorithms. It shows how appr...
Chapter
In the multi-agent system (MAS) context, the theories and practices of evolutionary computation (EC) have new implications, particularly with regard to engineering and shaping system behaviors. Thus, it is important that we consider the embodiment of EC in “real” agents, that is, agents that involve the real restrictions of time and space within MA...
Chapter
Publisher Summary New technologies for fighter aircrafts are being developed continuously. Often, aircraft engineers know a great deal about the aerodynamic performance of new fighter aircraft that exploit new technologies, even before a physical prototype is constructed or flown. Such aerodynamic knowledge is available from design principles, comp...
Article
Full-text available
For agent-based systems to reach their full potential, an important capability for individual agents is adaptation. An adaptive technique that is particularly well suited to the agentbased paradigm is provided by evolutionary computation (EC). EC systems have been shown to develop complex groups of coevolved structures. In fact, Holland's original...
Article
Full-text available
This paper reports the authors' ong oing experience with a system for discovering novel fig hter combat maneuvers, using a g enetics-based machine learning process, and combat simulation. In effect, theg enetic learning system in this application is taking the place of a test pilot, in discovering complex maneuvers from experience. Theg oal of this...
Article
Full-text available
There a re a number of common difficulties and open issues that pertain to the "traditional" LCS model. Many of these topics were central at The First International Workshop o n Learning Classifier Systems (Houston, Texas, 1992). Since the first workshop, several significant, theoretically-supported advances in LCS practice have addressed these iss...
Article
Full-text available
This paper reports the authors' ongoing experience with a system for discovering novel fighter combat maneuvers, using a genetics-based machine learning process, and combat simulation. In effect, the genetic learning system. in this application is taking the place of a test pilot, in discovering complex maneuvers from experience. The goal of this w...
Article
Full-text available
Reinforcement learning (RL) problems constitute an important class of learning and control problems faced by artificial intelligence systems. In these problems, one is faced with the task of providing control signals that maximize some measure of performance, usually taken over time, given feedback that is not in terms of the control signals themse...
Article
Full-text available
To understand the behaviour of search methods (including GAs), it is useful to understand the nature of the landscapes they search. What makes a landscape complex to search? Since there are an infinite number of landscapes, with an infinite number of characteristics, this is a di#cult question. Therefore, it is interesting to consider parameterised...
Article
Full-text available
In adaptive systems that involve large numbers of agents, emergent, global behaviours that arise from local agent interactions are a critical concept. In nature, such behaviours are central complex group behaviours that must arise from individuals that evolve selfishly. In artificial systems that mimic these adaptive, multi-agent models, understand...
Article
In adaptive systems that involve large numbers of entities, emergent, global behaviours that arise from localised interactions are a critical concept. Understanding and shaping emergence may be essential to such systems' success. To aid in this understanding, this paper introduces a measure gleaned from non-linear systems theory. The paper discusse...
Article
Full-text available
This paper carefully considers random landscapes related to Kauffman's NK model. In particular, it considers a superset of this model (the NKP model) recently suggested in the GA-analytic literature. Landscapes are exhaustively examined for both the distribution of local optima relative to the global optima, and for characteristics that would effec...
Conference Paper
Full-text available
This paper considers the application of XCS to the complex, real-world problem of mapping Boolean networks to technology-specific layout of field programmable gate arrays (FPGAs). The mapping is formulated as a temporal task, where the XCS's actions are to create blocks (based on an abstract Boolean network) that can be placed in the FPGA, one-at-a...