Improvisation in interactive music systems

This project investigates machine listening and improvisation in interactive music systems with the goal of improvising musically appropriate accompaniment to an audio stream in real-time. The input audio may be from a live musical ensemble, or playback of a recording for use by a DJ. I present a collection of robust techniques for machine listening in the context of Western popular dance music genres, and strategies of improvisation to allow for intuitive and musically salient interaction in live performance. The findings are embodied in a computational agent – the Jambot – capable of real-time musical improvisation in an ensemble setting. Conceptually the agent’s functionality is split into three domains: reception, analysis and generation. The project has resulted in novel techniques for addressing a range of issues in each of these domains. In the reception domain I present a novel suite of onset detection algorithms for real-time detection and classification of percussive onsets. This suite achieves reasonable discrimination between the kick, snare and hi-hat attacks of a standard drum-kit, with sufficiently low-latency to allow perceptually simultaneous triggering of accompaniment notes. The onset detection algorithms are designed to operate in the context of complex polyphonic audio. In the analysis domain I present novel beat-tracking and metre-induction algorithms that operate in real-time and are responsive to change in a live setting. I also present a novel analytic model of rhythm, based on musically salient features. This model informs the generation process, affording intuitive parametric control and allowing for the creation of a broad range of interesting rhythms. In the generation domain I present a novel improvisatory architecture drawing on theories of music perception, which provides a mechanism for the real-time generation of complementary accompaniment in an ensemble setting. All of these innovations have been combined into a computational agent – the Jambot, which is capable of producing improvised percussive musical accompaniment to an audio stream in real-time. I situate the architectural philosophy of the Jambot within contemporary debate regarding the nature of cognition and artificial intelligence, and argue for an approach to algorithmic improvisation that privileges the minimisation of cognitive dissonance in human-computer interaction. This thesis contains extensive written discussions of the Jambot and its component algorithms, along with some comparative analyses of aspects of its operation and aesthetic evaluations of its output. The accompanying CD contains the Jambot software, along with video documentation of experiments and performances conducted during the project.

Relative entropy rate based model selection for linear hybrid system filters of uncertain nonlinear systems

Hybrid system representations have been exploited in a number of challenging modelling situations, including situations where the original nonlinear dynamics are too complex (or too imprecisely known) to be directly filtered. Unfortunately, the question of how to best design suitable hybrid system models has not yet been fully addressed, particularly in the situations involving model uncertainty. This paper proposes a novel joint state-measurement relative entropy rate based approach for design of hybrid system filters in the presence of (parameterised) model uncertainty. We also present a design approach suitable for suboptimal hybrid system filters. The benefits of our proposed approaches are illustrated through design examples and simulation studies.

The acquisition of an artificial logographic script and bilingual working memory : Evidence for L1-specific orthographic processing skills transfer in Chinese-English bilinguals

Studies of orthographic skills transfer between languages focus mostly on working memory (WM) ability in alphabetic first language (L1) speakers when learning another, often alphabetically congruent, language. We report two studies that, instead, explored the transferability of L1 orthographic processing skills in WM in logographic-L1 and alphabetic-L1 speakers. English-French bilingual and English monolingual (alphabetic-L1) speakers, and Chinese-English (logographic-L1) speakers, learned a set of artificial logographs and associated meanings (Study 1). The logographs were used in WM tasks with and without concurrent articulatory or visuo-spatial suppression. The logographic-L1 bilinguals were markedly less affected by articulatory suppression than alphabetic-L1 monolinguals (who did not differ from their bilingual peers). Bilinguals overall were less affected by spatial interference, reflecting superior phonological processing skills or, conceivably, greater executive control. A comparison of span sizes for meaningful and meaningless logographs (Study 2) replicated these findings. However, the logographic-L1 bilinguals’ spans in L1 were measurably greater than those of their alphabetic-L1 (bilingual and monolingual) peers; a finding unaccounted for by faster articulation rates or differences in general intelligence. The overall pattern of results suggests an advantage (possibly perceptual) for logographic-L1 speakers, over and above the bilingual advantage also seen elsewhere in third language (L3) acquisition.

The state-of-the-art in personalized recommender systems for social networking

With the explosion of Web 2.0 application such as blogs, social and professional networks, and various other types of social media, the rich online information and various new sources of knowledge flood users and hence pose a great challenge in terms of information overload. It is critical to use intelligent agent software systems to assist users in finding the right information from an abundance of Web data. Recommender systems can help users deal with information overload problem efficiently by suggesting items (e.g., information and products) that match users’ personal interests. The recommender technology has been successfully employed in many applications such as recommending films, music, books, etc. The purpose of this report is to give an overview of existing technologies for building personalized recommender systems in social networking environment, to propose a research direction for addressing user profiling and cold start problems by exploiting user-generated content newly available in Web 2.0.

Personalised information gathering and recommender systems : techniques and trends

With the explosive growth of resources available through the Internet, information mismatching and overload have become a severe concern to users. Web users are commonly overwhelmed by huge volume of information and are faced with the challenge of finding the most relevant and reliable information in a timely manner. Personalised information gathering and recommender systems represent state-of-the-art tools for efficient selection of the most relevant and reliable information resources, and the interest in such systems has increased dramatically over the last few years. However, web personalization has not yet been well-exploited; difficulties arise while selecting resources through recommender systems from a technological and social perspective. Aiming to promote high quality research in order to overcome these challenges, this paper provides a comprehensive survey on the recent work and achievements in the areas of personalised web information gathering and recommender systems. The report covers concept-based techniques exploited in personalised information gathering and recommender systems.

Determining process model precision and generalization with weighted artificial negative events

Process mining encompasses the research area which is concerned with knowledge discovery from event logs. One common process mining task focuses on conformance checking, comparing discovered or designed process models with actual real-life behavior as captured in event logs in order to assess the “goodness” of the process model. This paper introduces a novel conformance checking method to measure how well a process model performs in terms of precision and generalization with respect to the actual executions of a process as recorded in an event log. Our approach differs from related work in the sense that we apply the concept of so-called weighted artificial negative events towards conformance checking, leading to more robust results, especially when dealing with less complete event logs that only contain a subset of all possible process execution behavior. In addition, our technique offers a novel way to estimate a process model’s ability to generalize. Existing literature has focused mainly on the fitness (recall) and precision (appropriateness) of process models, whereas generalization has been much more difficult to estimate. The described algorithms are implemented in a number of ProM plugins, and a Petri net conformance checking tool was developed to inspect process model conformance in a visual manner.

The use of artificial neural networks for identifying sustainable biodiesel feedstocks

Over the past few decades, biodiesel produced from oilseed crops and animal fat is receiving much attention as a renewable and sustainable alternative for automobile engine fuels, and particularly petroleum diesel. However, current biodiesel production is heavily dependent on edible oil feedstocks which are unlikely to be sustainable in the longer term due to the rising food prices and the concerns about automobile engine durability. Therefore, there is an urgent need for researchers to identify and develop sustainable biodiesel feedstocks which overcome the disadvantages of current ones. On the other hand, artificial neural network (ANN) modeling has been successfully used in recent years to gain new knowledge in various disciplines. The main goal of this article is to review recent literatures and assess the state of the art on the use of ANN as a modeling tool for future generation biodiesel feedstocks. Biodiesel feedstocks, production processes, chemical compositions, standards, physio-chemical properties and in-use performance are discussed. Limitations of current biodiesel feedstocks over future generation biodiesel feedstock have been identified. The application of ANN in modeling key biodiesel quality parameters and combustion performance in automobile engines is also discussed. This review has determined that ANN modeling has a high potential to contribute to the development of renewable energy systems by accelerating biodiesel research.

Ergonomic DHM systems - limitations and trends – a systematic literature review focused on the ‘future of ergonomics’

Digital Human Models (DHM) have been used for over 25 years. They have evolved from simple drawing templates, which are nowadays still used in architecture, to complex and Computer Aided Engineering (CAE) integrated design and analysis tools for various ergonomic tasks. DHM are most frequently used for applications in product design and production planning, with many successful implementations documented. DHM from other domains, as for example computer user interfaces, artificial intelligence, training and education, or the entertainment industry show that there is also an ongoing development towards a comprehensive understanding and holistic modeling of human behavior. While the development of DHM for the game sector has seen significant progress in recent years, advances of DHM in the area of ergonomics have been comparatively modest. As a consequence, we need to question if current DHM systems are fit for the design of future mobile work systems. So far it appears that DHM in Ergonomics are rather limited to some traditional applications. According to Dul et al. (2012), future characteristics of Human Factors and Ergonomics (HFE) can be assigned to six main trends: (1) global change of work systems, (2) cultural diversity, (3) ageing, (4) information and communication technology (ICT), (5) enhanced competiveness and the need for innovation, and; (6) sustainability and corporate social responsibility. Based on a literature review, we systematically investigate the capabilities of current ergonomic DHM systems versus the ‘Future of Ergonomics’ requirements. It is found that DHMs already provide broad functionality in support of trends (1) and (2), and more limited options in regards to trend (3). Today’s DHM provide access to a broad range of national and international databases for correct differentiation and characterization of anthropometry for global populations. Some DHM explicitly address social and cultural modeling of groups of people. In comparison, the trends of growing importance of ICT (4), the need for innovation (5) and sustainability (6) are addressed primarily from a hardware-oriented and engineering perspective and not reflected in DHM. This reflects a persistent separation between hardware design (engineering) and software design (information technology) in the view of DHM – a disconnection which needs to be urgently overcome in the era of software defined user interfaces and mobile devices. The design of a mobile ICT-device is discussed to exemplify the need for a comprehensive future DHM solution. Designing such mobile devices requires an approach that includes organizational aspects as well as technical and cognitive ergonomics. Multiple interrelationships between the different aspects result in a challenging setting for future DHM. In conclusion, the ‘Future of Ergonomics’ pose particular challenges for DHM in regards to the design of mobile work systems, and moreover mobile information access.

Vectors and methods for hairpin RNA and artificial microRNA-mediated gene silencing in plants

In plant cells, DICER-LIKE4 processes perfectly double-stranded RNA (dsRNA) into short interfering (si) RNAs, and DICER-LIKE1 generates micro (mi) RNAs from primary miRNA transcripts (pri-miRNA) that form fold-back structures of imperfectly dsRNA. Both si and miRNAs direct the endogenous endonuclease, ARGONAUTE1 to cleave complementary target single-stranded RNAs and either small RNA (sRNA)-directed pathway can be harnessed to silence genes in plants. A routine way of inducing and directing RNA silencing by siRNAs is to express self-complementary single-stranded hairpin RNA (hpRNA), in which the duplexed region has the same sequence as part of the target gene's mRNA. Artificial miRNA (amiRNA)-mediated silencing uses an endogenous pri-miRNA, in which the original miRNA/miRNA* sequence has been replaced with a sequence complementary to the new target gene. In this chapter, we describe the plasmid vector systems routinely used by our research group for the generation of either hpRNA-derived siRNAs or amiRNAs.

Combining recommender and reputation systems to produce better online advice

Although recommender systems and reputation systems have quite different theoretical and technical bases, both types of systems have the purpose of providing advice for decision making in e-commerce and online service environments. The similarity in purpose makes it natural to integrate both types of systems in order to produce better online advice, but their difference in theory and implementation makes the integration challenging. In this paper, we propose to use mappings to subjective opinions from values produced by recommender systems as well as from scores produced by reputation systems, and to combine the resulting opinions within the framework of subjective logic.

Developing Co-operating Legal Knowledge Based Systems

In attempting to build intelligent litigation support tools, we have moved beyond first generation, production rule legal expert systems. Our work supplements rule-based reasoning with case based reasoning and intelligent information retrieval. This research, specifies an approach to the case based retrieval problem which relies heavily on an extended object-oriented / rule-based system architecture that is supplemented with causal background information. Machine learning techniques and a distributed agent architecture are used to help simulate the reasoning process of lawyers. In this paper, we outline our implementation of the hybrid IKBALS II Rule Based Reasoning / Case Based Reasoning system. It makes extensive use of an automated case representation editor and background information.

The IKBALS project: multi-modal reasoning in legal knowledge based systems

In attempting to build intelligent litigation support tools, we have moved beyond first generation, production rule legal expert systems. Our work integrates rule based and case based reasoning with intelligent information retrieval. When using the case based reasoning methodology, or in our case the specialisation of case based retrieval, we need to be aware of how to retrieve relevant experience. Our research, in the legal domain, specifies an approach to the retrieval problem which relies heavily on an extended object oriented/rule based system architecture that is supplemented with causal background information. We use a distributed agent architecture to help support the reasoning process of lawyers. Our approach to integrating rule based reasoning, case based reasoning and case based retrieval is contrasted to the CABARET and PROLEXS architectures which rely on a centralised blackboard architecture. We discuss in detail how our various cooperating agents interact, and provide examples of the system at work. The IKBALS system uses a specialised induction algorithm to induce rules from cases. These rules are then used as indices during the case based retrieval process. Because we aim to build legal support tools which can be modified to suit various domains rather than single purpose legal expert systems, we focus on principles behind developing legal knowledge based systems. The original domain chosen was theAccident Compensation Act 1989 (Victoria, Australia), which relates to the provision of benefits for employees injured at work. For various reasons, which are indicated in the paper, we changed our domain to that ofCredit Act 1984 (Victoria, Australia). This Act regulates the provision of loans by financial institutions. The rule based part of our system which provides advice on the Credit Act has been commercially developed in conjunction with a legal firm. We indicate how this work has lead to the development of a methodology for constructing rule based legal knowledge based systems. We explain the process of integrating this existing commercial rule based system with the case base reasoning and retrieval architecture.

Reasoning paradigms in legal decision support systems

In this paper we discuss the strengths and weaknesses of a range of artificial intelligence approaches used in legal domains. Symbolic reasoning systems which rely on deductive, inductive and analogical reasoning are described and reviewed. The role of statistical reasoning in law is examined, and the use of neural networks analysed. There is discussion of architectures for, and examples of, systems which combine a number of these reasoning strategies. We conclude that to build intelligent legal decision support systems requires a range of reasoning strategies.

Near knowledge : inductive learning systems in law

Induction is an interesting model of legal reasoning, since it provides a method of capturing initial states of legal principles and rules, and adjusting these principles and rules over time as the law changes. In this article I explain how Artificial Intelligence-based inductive learning algorithms work, and show how they have been used in law to model legal domains. I identify some problems with implementations undertaken in law to date, and create a taxonomy of appropriate cases to use in legal inductive inferencing systems. I suggest that inductive learning algorithms have potential in modeling law, but that the artificial intelligence implementations to date are problematic. I argue that induction should be further investigated, since it has the potential to be an extremely useful mechanism for understanding legal domains.

Deductive, inductive and analogical reasoning in legal decision support systems

In this paper we provide an overview of a number of fundamental reasoning formalisms in artificial intelligence which can and have been used in modelling legal reasoning. We describe deduction, induction and analogical reasoning formalisms, and show how they can be used separately to model legal reasoning. We argue that these formalisms can be used together to model legal reasoning more accurately, and describe a number of attempts to integrate the approaches.