Given the unprecedented demand for mobile capacity beyond that available from the RF spectrum, it is natural to consider the infrared and visible light spectrum for future terrestrial wireless systems. Wireless systems using these parts of the electromagnetic spectrum could be classified as nmWave wireless communications system in relation to mmWave radio systems and both are being standardised in current 5G systems. TOWS, therefore, will provide a technically logical pathway to ensure that wireless systems are future-proof and that they can deliver the capacities that future data intensive services such as high definition (HD) video streaming, augmented reality, virtual reality and mixed reality will demand. Light based wireless communication systems will not be in competition with RF communications, but instead these systems follow a trend that has been witnessed in cellular communications over the last 30 years. Light based wireless communications simply adds new capacity - the available spectrum is 2600 times the RF spectrum. 6G and beyond promise increased wireless capacity to accommodate this growth in traffic in an increasingly congested spectrum, however action is required now to ensure UK leadership of the fast moving 6G field. Optical wireless (OW) opens new spectral bands with a bandwidth exceeding 540 THz using simple sources and detectors and can be simpler than cellular and WiFi with a significantly larger spectrum. It is the best choice of spectrum beyond millimetre waves, where unlike the THz spectrum (the other possible choice), OW avoids complex sources and detectors and has good indoor channel conditions. Optical signals, when used indoors, are confined to the environment in which they originate, which offers added security at the physical layer and the ability to re-use wavelengths in adjacent rooms, thus radically increasing capacity. Our vision is to develop and experimentally demonstrate multiuser Terabit/s optical wireless systems that offer capacities at least two orders of magnitude higher than the current planned 5G optical and radio wireless systems, with a roadmap to wireless systems that can offer up to four orders of magnitude higher capacity. There are four features of the proposed system which make possible such unprecedented capacities to enable this disruptive advance. Firstly, unlike visible light communications (VLC), we will exploit the infrared spectrum, this providing a solution to the light dimming problem associated with VLC, eliminating uplink VLC glare and thus supporting bidirectional communications. Secondly, to make possible much greater transmission capacities and multi-user, multi-cell operation, we will introduce a new type of LED-like steerable laser diode array, which does not suffer from the speckle impairments of conventional laser diodes while ensuring ultrahigh speed performance. Thirdly, with the added capacity, we will develop native OW multi-user systems to share the resources, these being adaptively directional to allow full coverage with reduced user and inter-cell interference and finally incorporate RF systems to allow seamless transition and facilitate overall network control, in essence to introduce software defined radio to optical wireless. This means that OW multi-user systems can readily be designed to allow very high aggregate capacities as beams can be controlled in a compact manner. We will develop advanced inter-cell coding and handover for our optical multi-user systems, this also allowing seamless handover with radio systems when required such as for resilience. We believe that this work, though challenging, is feasible as it will leverage existing skills and research within the consortium, which includes excellence in OW link design, advanced coding and modulation, optimised algorithms for front-haul and back-haul networking, expertise in surface emitting laser design and single photon avalanche detectors for ultra-sensitive detection.

Energy efficient processes are increasingly key priorities for ICT companies with attention being paid to both ecological and economic drivers. Although in some cases the use of ICT can be beneficial to the environment (for example by reducing journeys and introducing more efficient business processes), countries are becoming increasingly aware of the very large growth in energy consumption of telecommunications companies. For instance in 2007 BT consumed 0.7% of the UK's total electricity usage. In particular, the predicted future growth in the number of connected devices, and the internet bandwidth of an order of magnitude or two is not practical if it leads to a corresponding growth in energy consumption. Regulations may therefore come soon, particularly if Governments mandate moves towards carbon neutrality. Therefore the applicants believe that this proposal is of great importance in seeking to establish the current limits on ICT performance due to known environmental concerns and then develop new ICT techniques to provide enhanced performance. In particular they believe that substantial advances can be achieved through the innovative use of renewable sources and the development of new architectures, protocols, and algorithms operating on hardware which will itself allows significant reductions in energy consumption. This will represent a significant departure from accepted practices where ICT services are provided to meet the growing demand, without any regard for the energy consequences of relative location of supply and demand. In this project therefore, we propose innovatively to consider optimised dynamic placement of ICT services, taking account of varying energy costs at producer and consumer. Energy consumption in networks today is typically highly confined in switching and routing centres. Therefore in the project we will consider block transmission of data between centres chosen for optimum renewable energy supply as power transmission losses will often make the shipping of power to cities (data centres/switching nodes in cities) unattractive. Variable renewable sources such as solar and wind pose fresh challenges in ICT installations and network design, and hence this project will also look at innovative methods of flexible power consumption of block data routers to address this effect. We tackle the challenge along three axes: (i) We seek to design a new generation of ICT infrastructure architectures by addressing the optimisation problem of placing compute and communication resources between the producer and consumer, with the (time-varying) constraint of minimising energy costs. Here the architectures will leverage the new hardware becoming available to allow low energy operation. (ii) We seek to design new protocols and algorithms to enable communications systems to adapt their speed and power consumption according to both the user demand and energy availability. (iii) We build on recent advances in hardware which allow the block routing of data at greatly reduced energy levels over electronic techniques and determine hardware configurations (using on chip monitoring for the first time) to support these dynamic energy and communications needs. Here new network components will be developed, leveraging for example recent significant advances made on developing lower power routing hardware with routing power levels of approximately 1 mW/Gb/s for ns block switching times. In order to ensure success, different companies will engage their expertise: BT, Ericsson, Telecom New Zealand, Cisco and BBC will play a key role in supporting the development of the network architectures, provide experimental support and traffic traces, and aid standards development. Solarflare, Broadcom, Cisco and the BBC will support our protocol and intelligent traffic solutions. Avago, Broadcom and Oclaro will play a key role in the hardware development.

Dramatic progress has been made in the past few years in the field of photonic technologies, to complement those in electronic technologies which have enabled the vast advances in information processing capability. A plethora of new screen and projection display technologies have been developed, bringing higher resolution, lower power operation and enabling new ways of machine interaction. Advances in biophotonics have led to a large range of low cost products for personal healthcare. Advances in low cost communication technologies to rates now in excess of 10 Gb/s have caused transceiver unit price cost reductions from >$10,000 to less than $100 in a few years, and, in the last two years, large volume use of parallel photonics in computing has come about. Advances in polymers have made possible the formation of not just links but complete optical subsystems fully integrated within circuit boards, so that users can expect to commoditise bespoke photonics technology themselves without having to resort to specialist companies. These advances have set the scene for a major change in commercialisation activity where photonics and electronics will converge in a wide range of systems. Importantly, photonics will become a fundamental underpinning technology for a much greater range of users outside its conventional arena, who will in turn require those skilled in photonics to have a much greater degree of interdisciplinary training. In short, there is a need to educate and train researchers who have skills balanced across the fields of electronic and photonic hardware and software. The applicants are unaware of such capability currently.This Doctoral Training Centre (DTC) proposal therefore seeks to meet this important need, building upon the uniqueness of the Cambridge and UCL research activities that are already focussing on new types of displays based on polymer and holographic projection technology, the application of photonic communications to computing, personal information systems and indeed consumer products (via board-to-board, chip to chip and later on-chip interconnects), the increased use of photonics in industrial processing and manufacture, techniques for the low-cost roll-out of optical fibre to replace the copper network, the substitution of many conventional lighting products with photonic light sources and extensive application of photonics in medical diagnostics and personalised medicine. Many of these activities will increasingly rely on more advanced systems integration, and so the proposed DTC includes experts in computer systems and software. By drawing these complementary activities together, it is proposed to develop an advanced training programme to equip the next generation of very high calibre doctoral students with the required expertise, commercial and business skills and thus provide innovation opportunities for new systems in the future. It should be stressed that the DTC will provide a wide range of methods for learning for students, well beyond that conventionally available, so that they can gain the required skills. In addition to lectures and seminars, for example, there will be bespoke experimental coursework activities, reading clubs, roadmapping activities, secondments to collaborators and business planning courses.Photonics is likely to become much more embedded in other key sectors of the economy, so that the beneficiaries of the DTC are expected to include industries involved in printing, consumer electronics, computing, defence, energy, engineering, security, medicine and indeed systems companies providing information systems for example for financial, retail and medical industries. Such industries will be at the heart of the digital economy, energy, healthcare and nanotechnology fields. As a result, a key feature of the DTC will be a developed awareness in its cohorts of the breadth of opportunity available and a confidence that they can make impact therein.

Machine Learning (ML) already has a dramatic impact on our daily lives. ML developments in large language models and deep generative models cement that further. The recent explosion in ML, however, is built on the back of improved computer systems able to train and generate ever more powerful models. Systems design fundamentally defines ML performance and capability. This is true for Internet-scale ML and artificial intelligence (AI). Yet, more recently, it is especially evident in distributed, efficient, device-oriented, secure, personalised, privacy-preserving ML. UK strength in this fast developing area is dependent on a skilled R\&D workforce. Systems research and ML research are symbiotic. Current innovation in systems research is driven by the ubiquitous need for efficient and reliable ML. ML research, conversely, is steered by deployment capability and the economic and environmental impact of the resulting systems. Furthermore, systems research increasingly relies on ML methods to automate design, and ML research develops such methods. Major gains are made when the development of ML and systems are co-developed and co-optimized. This is relevant across a broad spectrum of industries: in-car systems, medical devices, mobile phones, sensor networks, condition monitoring systems, high-performance compute and high-frequency trading. Yet PhD training that brings together systems and ML is rare; research training is often siloed in the individual sub-disciplines. Instead, we need researchers trained in both fields and experienced in working across them. Hence: The ML Systems CDT will train a new type of student -- the ML-systems researcher. The ML Systems researcher is critically capable in both fields, and has collaborative research experience across the systems-ML stack. An example concretises this. A company is developing and deploying wearable body monitors. Effective models must be learnt on collected data, but data must be privacy preserving and bandwidth minimized. This is then personalised to each individual, adaptable to circumstance while being battery efficient and not connection dependent. To manage such a project requires knowledge of effective data-efficient ML signal analysis methods, designed and optimized for low-power hardware, itself tailored for the purpose through ML optimization methods. Knowledge of personalisation methods and the payoffs of privacy preserving methods vitally complement this. The societal impact, e.g.\ on those who might be obsessive about their medical state must also be considered, and will impact development. This CDT will train individuals with cross-cutting capability in all these components. Students must have broad understanding of different hardware designs, different platforms, different environments, different models, and different goals beyond their immediate research focus. This makes a cohort-based CDT vital. Standard PhD training in ML systems can result in research focus on a single ML technique and a single system. The CDT treats ML Systems as a holistic discipline. Cohort interaction, and integration gives students real experience across multiple systems, approaches and methodologies. Furthermore students will join together to contribute to a unified toolkit for the ML-Systems stack, and make use of others' contributions to that toolkit. On leaving the CDT, our graduates will understand fully where to focus resources to best improve a company's real-world ML development - whether that be at the ML-algorithm level, the hardware level, the compiler, level or even the legal level. They will be able to evaluate work at every level. We expect our graduates to be the leading team managers in real-world cutting-edge company ML.

Humans are highly adaptable, and speech is our natural medium for informal communication. When communicating, we continuously adjust to other people, to the situation, and to the environment, using previously acquired knowledge to make this adaptation seem almost instantaneous. Humans generalise, enabling efficient communication in unfamiliar situations and rapid adaptation to new speakers or listeners. Current speech technology works well for certain controlled tasks and domains, but is far from natural, a consequence of its limited ability to acquire knowledge about people or situations, to adapt, and to generalise. This accounts for the uneasy public reaction to speech-driven systems. For example, text-to-speech synthesis can be as intelligible as human speech, but lacks expression and is not perceived as natural. Similarly, the accuracy of speech recognition systems can collapse if the acoustic environment or task domain changes, conditions which a human listener would handle easily. Research approaches to these problems have hitherto been piecemeal and as a result progress has been patchy. In contrast NST will focus on the integrated theoretical development of new joint models for speech recognition and synthesis. These models will allow us to incorporate knowledge about the speakers, the environment, the communication context and awareness of the task, and will learn and adapt from real world data in an online, unsupervised manner. This theoretical unification is already underway within the NST labs and, combined with our record of turning theory into practical state-of-the-art applications, will enable us to bring a naturalness to speech technology that is not currently attainable.The NST programme will yield technology which (1) approaches human adaptability to new communication situations, (2) is capable of personalised communication, and (3) takes account of speaker intention and expressiveness in speech recognition and synthesis. This is an ambitious vision. Its success will be measured in terms of how the theoretical development reshapes the field over the next decade, the takeup of the software systems that we shall develop, and through the impact of our exemplar interactive applications.We shall establish a strong User Group to maximise the impact of the project, with a members concerned with clinical applications, as well as more general speech technology. Members of the User Group include Toshiba, EADS Innovation Works, Cisco, Barnsley Hospital NHS Foundation Trust, and the Euan MacDonald Centre for MND Research. An important interaction with the User Group will be validating our systems on their data and tasks, discussed at an annual user workshop.