Within the next few years the number of devices connected to each other and the Internet will outnumber humans by almost 5:1. These connected devices will underpin everything from healthcare to transport to energy and manufacturing. At the same time, this growth is not just in the number or variety of devices, but also in the ways they communicate and share information with each other, building hyper-connected cyber-physical infrastructures that span most aspects of people's lives. For the UK to maximise the socio-economic benefits from this revolutionary change we need to address the myriad trust, identity, privacy and security issues raised by such large, interconnected infrastructures. Solutions to many of these issues have previously only been developed and tested on systems orders of magnitude less complex in the hope they would 'scale up'. However, the rapid development and implementation of hyper-connected infrastructures means that we need to address these challenges at scale since the issues and the complexity only become apparent when all the different elements are in place. There is already a shortage of highly skilled people to tackle these challenges in today's systems with latest estimates noting a shortfall of 1.8M by 2022. With an estimated 80Bn malicious scans and 780K records lost daily due to security and privacy breaches, there is an urgent need for future leaders capable of developing innovative solutions that will keep society one step ahead of malicious actors intent on compromising security, privacy and identity and hence eroding trust in infrastructures. The Centre for Doctoral Training (CDT) 'Trust, Identity, Privacy and Security - at scale' (TIPS-at-Scale) will tackle this by training a new generation of interdisciplinary research leaders. We will do this by educating PhD students in both the technical skills needed to study and analyse TIPS-at-scale, while simultaneously studying how to understand the challenges as fundamentally human too. The training involves close involvement with industry and practitioners who have played a key role in co-creating the programme and, uniquely, responsible innovation. The implementation of the training is novel due to its 'at scale' focus on TIPS that contextualises students' learning using relevant real-world, global problems revealed through project work, external speakers, industry/international internships/placements and masterclasses. The CDT will enrol ten students per year for a 4-year programme. The first year will involve a series of taught modules on the technical and human aspects of TIPS-at-scale. There will also be an introductory Induction Residential Week, and regular masterclasses by leading academics and industry figures, including delivery at industrial facilities. The students will also undertake placements in industry and research groups to gain hands-on understanding of TIPS-at-scale research problems. They will then continue working with stakeholders in industry, academia and government to develop a research proposal for their final three years, as well as undertake internships each year in industry and international research centres. Their interdisciplinary knowledge will continue to expand through masterclasses and they will develop a deep appreciation of real-world TIPS-at-scale issues through experimentation on state-of-the-art testbed facilities and labs at the universities of Bristol and Bath, industry and a city-wide testbed: Bristol-is-Open. Students will also work with innovation centres in Bath and Bristol to develop novel, interdisciplinary solutions to challenging TIPS-at-scale problems as part of Responsible Innovation Challenges. These and other mechanisms will ensure that TIPS-at-Scale graduates will lead the way in tackling the trust, identity, privacy and security challenges in future large, massively connected infrastructures and will do so in a way that considers wider sosocial responsibility.

Recent advances in server and network virtualisation have given rise to the Infrastructure-as-a-Service paradigm where businesses can lease resources from cloud datacentre operators, thus enabling the outsourcing of ICT. Such businesses can themselves be application and service providers who act as tenants of a shared data centre infrastructure. The tenants resize their ICT footprint through the pay-as-you-go pricing model, thereby maintaining low capital (and operational) expenditure and increasing their profit margin. This infrastructural abstraction allows tenants to focus solely on their business delivery model while leaving the infrastructure maintenance to the operators. However, the resulting lack of visibility to the dynamic state of the underlying infrastructure can immensely hurt the services of the tenants when its performance fluctuates in short timescales. This prohibits the more pervasive migration of businesses to the cloud who are instead forced to maintain their own, in-house infrastructures. Adding to the problem, security risks are more acute in the cloud. Attackers can leverage cloud servers to launch DDoS attacks to other tenants or faster portscan to identify vulnerable services. Especially tenants are completely excluded from detecting security threats and from taking remedial action autonomously as the incidents unfold. Vulnerable services can end up consuming immense amounts of compute and network resources, leading to unsustainable bills for tenants who ultimately may have to retreat their services from the cloud. Existing measurement and monitoring approaches are inadequate because they are architected specifically for accounting, traffic engineering or offline debugging. Measurements from these approaches provide no clue on whether an application suffers self-induced congestion or cyber-attacks, there are some other offending flows/applications, or unacceptable latencies are due to long queueing delay at certain switch or application components, and how many flows are impacted by them. While addressing these problems itself is important to cloud operators, doing so in a timely fashion is often simply impossible because software and hardware updates take time and new pathological traffic patterns may arise as applications evolve. The overarching goal of this project is to design and develop a native Network Measurement-as-a-Service (NMaaS) framework that will allow tenants to express their measurement needs, and to subsequently synthesise the corresponding complex service-level performance functions out of simple monitoring primitives. The required primitive measurement components will be dynamically and transparently instantiated when and where required throughout the infrastructure, exploiting the temporal available capacity of servers and network nodes. In particular, we aim to: - devise novel server and switch instrumentation capabilities for traffic monitoring and make them as a native part of an underlying infrastructure so that they can support diverse measurement functions while alleviating measurement errors and uncertainties - develop a network-wide, centrally-orchestrated algorithm for the synthesis of complex metrics through the optimal placement of server-based and switch-based measurement functions in virtual and physical network components - design and develop measurement requirement description APIs to parse high-level measurement specifications issued by tenants and transform them into low-level measurement indicators. Ultimately, we aim to demonstrate that the proposed framework will contribute significantly in maintaining the desired application performance while at the same time improving the utilisation of cloud resources. Given that the cloud is still a rapidly growing global business, we anticipate that the research outcome will greatly benefit the wider IT industry.

This project is a direct response to significant changes taking place in the domain of computing and the arts. Recent developments in Artificial Intelligence and Machine Learning are leading to a revolution in how music and art is being created by researchers (Broad and Grierson, 2016). However, this technology has not yet been integrated into software aimed at creatives. Due to the complexities of machine learning, and the lack of usable tools, such approaches are only usable by experts. In order to address this, we will create new, user-friendly technologies that enable the lay user - composers as well as amateur musicians - to understand and apply these new computational techniques in their own creative work. The potential for machine learning to support creative activity is increasing at a significant rate, both in terms of creative understanding and potential applications. Emerging work in the field of music and sound generation extends from musical robots to generative apps, and from advanced machine listening to devices that can compose in any given style. By leveraging the internet as a live software ecosystem, the proposed project examines how such technology can best reach artists, and live up to its potential to fundamentally change creative practice in the field. Rather than focussing on the computer as an original creator, we will create platforms where the newest techniques can be used by artists as part of their day-to-day creative practices. Current research in artificial intelligence, and in particular machine learning, have led to an incredible leap forward in the performance of AI systems in areas such as speech and image recognition (Cortana, Siri etc.). Google and others have demonstrated how these approaches can be used for creative purposes, including the generation of speech and music (DeepMinds's WaveNet and Google's Magenta), images (Deep Dream) and game intelligence (DeepMind's AlphaGo). The investigators in this project have been using Deep Learning, Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory Networks (LSTMs), and other approaches to develop intelligent systems that can be used by artists to create sound and music. We are already among the first in the world to create reusable software that can 'listen' to large amounts of sound recordings, and use these as examples to create entirely new recordings at the level of audio. Our systems produce outcomes that out-perform many other previously funded research outputs in these areas. In this three-year project, we will develop and disseminate creative systems that can be used by musicians and artists in the creation of entirely new music and sound. We will show how such approaches can affect the future of other forms of media, such as film and the visual arts. We will do so by developing a creative platform, using the most accessible public forum available: the World Wide Web. We will achieve this through development of a high level live coding language for novice users, with simplified metaphors for the understanding of complex techniques including deep learning. We will also release the machine learning libraries we create for more advanced users who want to use machine learning technology as part of their creative tools. The project will involve end-users throughout, incorporating graduate students, professional artists, and participants in online learning environments. We will disseminate our work early, gaining the essential feedback required to deliver a solid final product and outcome. The efficacy of such techniques has been demonstrated with systems such as Sonic Pi and Ixi Lang, within a research domain already supported by the AHRC through the Live Coding Network (AH/L007266/1), and by EC in the H2020 project, RAPID-MIX. Finally, this research will strongly contribute to dialogues surrounding the future of music and the arts, consolidating the UK's leadership in these fields.

The strategic objective of this platform grant is to underpin Audio-Visual Media Research within the Centre for Vision, Speech and Signal Processing (CVSSP) to pursue fundamental research combining internationally leading expertise in understanding of real-world audio and visual data, and to transfer this capability to impact new application domains. Our goal is to pioneer new technologies which impact directly on industry practice in healthcare, sports, retail, communication, entertainment and training. This builds on CVSSP's unique track-record of world-leading research in both audio and visual machine perception which has enabled ground-breaking technology exploited by UK industry. The strategic contribution and international standing of the centres research in audio and visual media has been recognised by EPSRC through two previous platform grant awards (2003-14) and two programme grant awards in 2013 and 2015. Platform Grant funding is requested to reinforce the critical mass of expertise and knowledge of specialist facilities required to contribute advance in both fundamental understanding and pioneering new technology. In particular this Platform Grant will catalyse advances in multi-sensory machine perception building on the Centre's unique strengths in audio and vision. Key experienced post-doctoral researchers have specialist knowledge and practical know-how, which is an important resource for training new researchers and for maintaining cutting edge research using state-of-the-art facilities. Strategically the Platform Grant will build on recent independent advances in audio and visual scene analysis to lead multi-sensory understanding and modelling of real-world scenes. Research advances will provide the foundation for UK industry to lead the development of technologies ranging from intelligent sensing for healthcare and assisted living to immersive entertainment production. Platform Grant funding will also strengthen CVSSP's international collaboration with leading groups world-wide through extended research secondments US (Washington, USC), Asia (Tsinghua, Tianjin, Kyoto, Tokyo, KAUST) and Europe (INRIA, MPI, Fraunhofer, ETH, EPFL, KTH, CTU, UPF).

Testing is a crucial part of any software development process. Testing is also very expensive: Common estimations list the effort of software testing at 50% of the average budget. Recent studies suggest that 77% of the time that software developers spend with testing is used for reading tests. Tests are read when they are generated, when they are updated, fixed, or refactored, when they serve as API usage examples and specification, or during debugging. Reading and understanding tests can be challenging, and evidence suggests that, despite the popularity of unit testing frameworks and test-driven development, the majority of software developers do not practice testing actively. Automatically generated tests tend to be particularly unreadable, severely inhibiting the widespread use of automated test generation in practice. The effects of insufficient testing can be dramatic, with large economic damage, and the potential to harm people relying on software in safety critical applications. Our proposed solution to address this problem is to improve the effectiveness and efficiency of testing by improving the readability of tests. We will investigate which syntactic and semantic aspects make tests readable, such that we can make readability measurable by modelling it. This, in turn, will allow us to provide techniques that guide manual or automatic improvement of the readability of software tests. This is made possible by a unique combination of machine learning, crowd sourcing, and search-based testing techniques. The GReaTest project will provide tools to developers that help them to identify readability problems, to automatically improve readability, and to automatically generate readability optimised test suites. The importance of readability and the usefulness of readability improvement will be evaluated with a range of empirical studies in conjunction with our industrial collaborators Microsoft, Google, and Barclays, investigating the relation of test readability to fault finding effectiveness, developer productivity, and software quality. Automated analysis and optimisation of test readability is novel, and traditional analyses only focused on easily measurable program aspects, such as code coverage. Improving readability of software tests has a direct impact on industry, where testing is a major economic and technical factor: More readable tests will reduce the costs of testing and increase effectiveness, thus improving software quality. Readability optimisation will be a key enabler for automated test generation in practice. Once readability of software tests is understood, this opens the doors to a new research direction on analysis and improvement of other software artefacts based on human understanding and performance.