Search

Recent technological innovations, such as artificial intelligence (AI) and robotics, have attracted much scholarly attention, though largely in respect of their effect on the labour market. The thesis documents the indirect impact of technological innovations on the mental health and voting decisions of workers in the United States. The first paper investigates whether AI can be served as a general-purpose technology, examining comprehensive data of AI patents. It emerges that the stocks of AI related patents, such as deep learning and logistics systems, grow more quickly than robotics ones. AI applications are more popular in the chemical and medical sectors than those of computer science and electronics. The findings of the first paper indicate the potential of AI to serve as a general-purpose technology. Thus, policy makers should consider the readiness of our labour market, political and educational institutions to adjust to the introduction of AI. The second paper empirically explores the impact of industrial robots on voting outcomes in the U.S. presidential and congressional elections. It studies industrial robot and election data from the United States; the endogeneity issue is addressed by using the European robot usage as the instrument of U.S. robot exposure. It is deduced that industrial robots significantly influence voting behaviour. As such, politicians might seek to mitigate the sway of robots to safeguard political stability. The third paper empirically assesses how industrial robots affect the mental health of U.S. workers. Statistics of mortalities caused by drug and alcohol abuse, among others, are used as indicators of worker mental health. Industrial robots are found to exert a stark, negative effect on the mental health of workers, reflected by a rising drug and alcohol-induced mortality rate, for example. In other words, increasing robot usage leads to deterioration in mental health. Companies seeking to widen the participation of robotics should take this into account. Likewise, policy makers should provide better protection and welfare for those workers at risk of job loss due to technological advancement. Overall, we confirm the potential for AI to reach far and wide across our economy. Furthermore, the ability of industrial robots to affect election results and the mental health of individuals is evidenced. Accordingly, it is pivotal that policy makers are prepared for the expansion of AI and able to identify and nullify the possible risks.

This study contributes to practical discussions on the composition of dynamic music for video games from the composer’s perspective. Creating greater levels of immersion in players is used as a justification for the proposals of the thesis. It lays down foundational aesthetic elements in order to proceed with a logical methodology. The aim of this paper is to build upon, and further hybridise, two techniques used by composers and by video game designers to increase further the reactive agility and memorability of the music for the player. Each chapter of this paper explores a different technique for joining two (possibly disparate) types of gameplay, or gamestates, with appropriate continuous music. In each, I discuss a particular musical engine capable of implementing continuous music. Chapter One will discuss a branching-music engine, which uses a precomposed musical mosaic (or musical pixels) to create a linear score with the potential to diverge at appropriate moments accompanying onscreen action. I use the case study of the Final Fantasy battle system to show how the implementation of a branching-music engine could assist in maintaining the continuity of gameplay experience that current disjointed scores, which appear in many games, create. To aid this argument I have implemented a branching-music engine, using the graphical object oriented programming environment MaxMSP, in the style of the battle music composed by Nobuo Uematsu, the composer of the early Final Fantasy series. The reader can find this in the accompanying demonstrations patch. In Chapter Two I consider how a generative-music engine can also implement a continuous music and also address some of the limitations of the branching-music engine. Further I describe a technique for an effective generative music for video games that creates musical ‘personalities’ that can mimic a particular style of music for a limited period of time. Crucially, this engine is able to transition between any two personalities to create musical coincidence with the game. GMGEn (Game Music Generation Engine) is a program I have created in MaxMSP to act as an example of this concept. GMGEn is available in the Demonstrations_Application. Chapter Three will discuss potential limitations of the branching music engine described in Chapter One and the generative music engine described in Chapter Two, and highlights how these issues can be solved by way of a third engine, which hybridises both. As this engine has an indeterminate musical state it is termed the intermittent-music engine. I go on to discuss the implementation of this engine in two different game scenarios and how emergent structures of this music will appear. The final outcome is to formulate a new compositional approach delivering dynamic music, which accompanies the onscreen action with greater agility than currently present in the field, increasing the memorability and therefore the immersive effect of the video-game music.

This portfolio of compositions uses novel biometrics from gestural interfaces called Myo armbands to compose five pieces of electroacoustic music within a continuum of spatial environments, ranging from the real to the virtual. The portfolio achieves this by processing biometric data from the Myo armbands within developed software (using Max 8), applying machine learning to such data (via Wekinator) and mapping predictive outputs to audiovisual materials across the spatial continuum. The results of this approach show how unique compositional affordances can be created when using biometrics within different spatial environments and how implementing machine learning methods can help to create such opportunities. The outcomes are presented via five composed portfolio works, three published papers within the scientific community, and numerous music performances navigating the overlapping areas of music composition and computer science. By using novel biometrics to compose music within a continuum of space, this interdisciplinary research topic contributes original knowledge to both the fields of music composition and computer science (human-computer interaction).

Small calibre projectiles are spin-stabilised to increase ballistic stability, often at high frequencies. Due to hardware limitations, conventional actuators and meth ods are unable to provide satisfactory control at such high frequencies. With the reduced volume for control hardware and increased financial cost, incorporating traditional guid ance methods into small-calibre projectiles is inherently difficult. This work presents a novel method of projectile control which addresses these issues and conducts a systems level analysis of the underlying actuation mechanism. The design is shown to be a viable alternative to traditional control methods, Firstly, a 7 Degree-of-Freedom (DoF) dynamic model is created for dual-spin pro jectiles, including aerodynamic coefficients. The stability of dual-spin projectiles, gov erned by the gyroscopic and dynamic stability factors is given, discussed and unified across available literature. The model is implemented in a Matlab/Simulink simulation environ ment, which is in turn validated against a range of academic literature and experimental test data. The novel design and fundamental operating principle are presented. The actuation mechanism (AM) is then mathematically formulated from both a velocity change (∆V ) and a lateral acceleration (a˜) perspective. A set of axioms are declared and verified using the 7-DoF model, showing that the inherently discrete system behaviour can be controlled continuously via these control variables, ∆V or a˜. Control state switching is simplified to be instantaneous, then expanded to be generically characterised by an arbitrarily complex mathematical function. A detailed investigation, parametric analysis and sensitivity study is undertaken to understand the system behaviour. A Monte Carlo procedure is described, which is used to compare the correction cap abilities of different guidance laws (GLs). A bespoke Zero-Effort-Miss (ZEM) based GLis synthesised from the mathematical formulation of the AM, with innately more know ledge of the system behaviour, which allows superior error correction. This bespoke GL is discussed in detail, a parametric study is undertaken, and both the GL parameters and PID controller gains are optimised using a genetic algorithm. Artificial Intelligence (AI) Reinforcement learning methods are used to emulate a GL, as well as controlling the AM and operating as a GL, simultaneously. The novel GLs are compared against a traditional proportional navigation GL in a nominal system and all GLs were able to control the AMs, reducing the miss distance to a satisfactory margin. The ZEM-based GL provided superior correction to the AI GL, which in turn provided superior correction over proportional navigation. Example CAD models are shown, and the stability analysis is conducted on the geometry. The CAD model is then used in CFD simulations to determine aerodynamic coefficients for use in the 7-DoF dynamic model. The novel control method was able to reduce the 95% dispersion diameter of a traditional ballistic 7.62mm projectile from 70mm to 33mm. Statistical data analysis showed there was no significant correlation or bias present in either the nominal or 7-DoF dispersion patterns. This project is co-sponsored by BAE Systems and ESPRC (ref. 1700064). The con tents of this thesis are covered by patent applications GB2011850.1, GB 2106035.5 and EP 20275128.5. Two papers are currently published (DOI: 10.1016/j.dt.2019.06.003, the second DOI is pending) and one is undergoing peer review.

Probabilistic programming is a paradigm that enables us to efficiently write probabilistic models as program code that we can sample, infer underlying parameters and predict outcomes based on complete or incomplete observations. Naturally, stochastic simulators, a special sub-class of simulators containing random variables, internal inference procedures, and the simulation of observations, are structurally rich probabilistic models. However, most simulators are not written in the probabilistic programming paradigm, as they are written in arbitrary programming code. This means that it is challenging to automatically update the variables in these simulators to account for observations from conducted experiments, which limits the simulators' use. Furthermore, there are two components to a probabilistic programming system i) the language and compilation procedure, ii) the inference procedures. These components can limit our ability to compile particular classes of probabilistic models, such as models that contain mixtures of parameter types, due to restrictions in the expressiveness of the language. Restrictions in the expressivity of the language can also inhibit our ability to generate efficient inferences, as this naturally influences the design of the probabilistic programming system and the set of available inference backends. Creating probabilistic programming systems that are expressive enough for different probabilistic models leads to the creation of many different probabilistic programming systems, which is inefficient - it would be more efficient if we could repurpose existing probabilistic programming systems. In this thesis, we develop three pieces of original work through four papers. The first piece of work describes how to extend differentiable first-order probabilistic programming systems to perform statistically correct and computationally efficient inference on models with mixtures of continuous and non-continuous parameters, without having to modify the underlying language, or develop an entirely new probabilistic programming system. The second describes how to translate real-world stochastic simulators written in arbitrary program languages to probabilistic programming systems. And finally, in the third piece of work, we develop two new Bayesian inference schemes to make inference more computationally and statistically efficient in nested models, models that contain probabilistic programs, within probabilistic programs, which arise in many real-world stochastic simulators.

In the mid-1950s, researchers in the United States melded formal theories of problem solving and intelligence with another powerful new tool for control: the electronic digital computer. Several branches of western mathematical science emerged from this nexus, including computer science (1960s–), data science (1990s–) and artificial intelligence (AI). This thesis offers an account of the origins and politics of AI in the mid-twentieth century United States, which focuses on its imbrications in systems of societal control. In an effort to denaturalize the power relations upon which the field came into being, I situate AI’s canonical origin story in relation to the structural and intellectual priorities of the U.S. military and American industry during the Cold War, circa 1952 to 1961. This thesis offers a detailed and comparative account of the early careers, research interests, and key outputs of four researchers often credited with laying the foundations for AI and machine learning—Herbert A. Simon, Frank Rosenblatt, John McCarthy and Marvin Minsky. It chronicles the distinct ways in which each sought to formalise and simulate human mental behaviour using digital electronic computers. Rather than assess their contributions as discontinuous with what came before, as in mythologies of AI's genesis, I establish continuities with, and borrowings from, management science and operations research (Simon), Hayekian economics and instrumentalist statistics (Rosenblatt), automatic coding techniques and pedagogy (McCarthy), and cybernetics (Minsky), along with the broadscale mobilization of Cold War-era civilian-led military science generally. I assess how Minsky’s 1961 paper 'Steps Toward Artificial Intelligence' simultaneously consolidated and obscured these entanglements as it set in motion an initial research agenda for AI in the following two decades. I argue that mind-computer metaphors, and research in complex information processing generally, played an important role in normalizing the small- and large-scale structuring of social behaviour using mathematics in the United States from the second half of the twentieth century onward.

The foreign exchange (Forex) market, as one of the most important financial markets in the globe, has attracted many investors. In order to support forex traders’ trading decisions, accurately predicting the forex prices has continued to be a popular but challenging topic. Due to the high complexity of the forex market, it is always a question of how effective the forex prediction could be. With the rapid development on machine learning in the last decades, deep learning has been applied successfully to many areas including the forex market. Consequently, numerous research papers have been published, which aim to improve the accuracy of forex prediction. The Long Short-Term Memory (LSTM) neural network, a kind of artificial neural network, has been widely used, which is specially designed to analyse time series data. Due to its strong learning capability, the LSTM neural network has now been used to predict complex forex trading based on historical data. However, there is a lack of an authoritative and commonly accepted guidance on how to conduct proper forex predictions by using LSTM. The application of deep learning to financial forecasting is still in a developing stage. This research aims to investigate the feasibility of applying deep learning, particularly the LSTM neural network to the foreign exchange market and to enhance the prediction accuracy via improved LSTM algorithms. In this thesis, all of the fundamental and technical features related to forex trading have been collected and analysed comprehensively. The influential features are then selected to be used as the inputs for forex prediction. Based on these inputs, a LSTM is specifically built to predict the trends of forex prices, which are identified as a suitable prediction target for forex traders. Notably, a new validation method is also introduced to overcome the problems in the traditional time series validation methods. Furthermore, a novel LSTM algorithm using hybrid activation functions in the same hidden layer is proposed to improve the prediction accuracy for forex trend predictions. Extensive experiments have been conducted and the experimental results have shown that the performance of the LSTM with hybrid activation functions has outperformed that of the standard LSTM. The generasalibility of the hybrid activation functions based LSTM has also been proved by its successful applications to different ANNs (e.g., RNNs) and datasets.