What do you ask a quantum computer?

“€1 billion for first quantum computer: race between Europe and US has really started” and “Quantum computers are the calculators of the future. They can solve complex problems with almost unlimited computing power” - Martijn van Calmthout, de Volkskrant, 27 November 2017. Kees Vuik, professor of numerical analysis and director of the TU Delft Institute for Computational Science and Engineering, and Matthias Möller, assistant professor, are wondering how it will be possible to deploy quantum computers to solve engineering problems fifteen years from now. Physicists think that we will be able to do everything once we have a quantum computer. But HOW, that is what we still need to figure out. Kees Vuik Simulation-driven modelling and optimisation of challenging engineering problems, such as the design of a ‘greener’ aeroplane, requires a lot of manual work and human intuition and cannot be fully automatized with today’s computing technologies. This is due to the thousands of design parameters that need to be optimised simultaneously, which exceeds the capabilities of today’s computing hardware. This applies both to the computation speed of the processors and the reading from and writing to memory, known as the ‘memory wall’. What we in Delft are therefore trying to do is to make our numerical methods as compatible as possible with the hardware. This means translating the engineering problem, or parts of it, into mathematical calculation procedure in such a way that specific hardware (such as GPUs or FPGAs) deliver extra computation speed. Preferably we want to read from memory once, then efficiently compute a great deal, and then write to memory once again. The quantum computer is the hardware of the future, and with its special parallel computing power (quantum parallelism) it can solve certain calculations at speeds exponentially higher than the best current and future supercomputers. That is to say, in several days rather than many decades. However, to exploit this enormous potential we have to reformulate our mathematical calculation procedure completely from scratch. For example, the number of input and output parameters for a quantum algorithm is limited by the number of qubits (quantum bits). In the coming years, these will be several dozen as a maximum, which in a current computer translates to just a single double-precision floating point number unless smarter data encoding schemes are developed. With a classic computer you provide many thousands, or hundreds of thousands, of input parameters, in the pre-processing step, and calculate, for example, the aerodynamic flow field around an aeroplane wing at millions of grid points. You then convert this, in the post-processing step, into a small number of relevant parameters, such as lift and drag coefficient telling you the quality of your design. This traditional approach is likely to fail on a quantum computer for various reasons. First of all, quantum computers cannot copy data unlike classic ones can do. Next, any storage of data in classical sense destroys the superposition phenomenon, which is giving the quantum computer its exceptional computing power. A quantum computer should therefore perform the modelling, simulation and optimisation steps for your specific engineering problem in one shot, with only a couple of input parameters. Furthermore, due to the very limited number of qubits, you have to ask a quantum computer a question for which the calculation is complex, but for which the answer is a single number or – preferably – yes or no. For example: ‘Is wing A more efficient than wing B?’ This requires a different mindset. And it also implies – although we still cannot oversee this – a completely different way of doing mathematics. By this time next year, we hope to be able to solve a linear system with a 5x5 matrix using a quantum computer. In fifteen years’ time, when we will have the use of several thousands of qubits, we want to be able to do this at a scale which is useful in the practical field. The concept has to be ready then, so we are already starting on it now. The insights we obtain could result in a reconsideration of the architecture of the quantum computer itself. At the same time, they could also result in benefits in terms of use of the computer hardware currently available. It will not be possible to translate everything we want to model into a quantum algorithm. We predict a multi-purpose computer in the future, probably as a cloud service, in which parts of the modelling problem will be optimised for computing using a quantum algorithm, and other parts for a GPU, for example. We want to prepare our field and our students to be able to solve engineering problems with maximum efficiency using such a multi-purpose computer. Text: Martijn Engelsman | Photo: Mark Prins

Robots, big data and the internet of things

“Robots taking care of the elderly. It’s the nightmare vision of many now that there are more ways than ever for machine assistance” and “The rapid evolution of robots raises many moral dilemma’s.” – Ad Ermstrang, ‘Robots als mensen: de nieuwe torenbouw van Babel’, Reformatorisch Dagblad, 8 januari 2018. In 2050, when meeting a robot, you actually meet the physical manifestation of big data Ir. Dr. C.J.M. (Chris) Verhoeven, associate professor in the micro-electronics department and theme-leader of swarm-robotics at the TU Delft Robotics Institute foresees a robotic future with a clear vision as to their intelligence and autonomy. It is a major societal need to be surrounded by autonomous devices and robots will fulfil that need. The rise of the robots is, however, interlocked with big data and the internet of things. The robotic trinity is the foundation of a society in which we will no longer be owners of hardware but of behaviour. The moment you step into an autonomous car in Spain, that car will be your car. Your desired driving style will be downloaded from the spacecloud, a swarm of nano-satellites packed with memory. The health support robot, the robot dog and the thermostat at your vacation home will also react to your existing cloud data, eliminating the need for any further commands. This need for autonomy has subtle but important characteristics. The roman horseman had no desire to control the horse’s muscles. The rider commands, the horse runs. Together they are a well-oiled, fast, flexible and safe unit. The autonomic car will similarly be in control of how to drive and which route to take, with us in command. Autonomous behaviour, under our supervision and responsibility. At several levels of intelligence mankind has a need to be surrounded by autonomous devices A further characteristic of autonomous robots is the absence of centralized control, as that would enable malicious interventions. Robots communicate with the cloud and big data provides them with powerful capabilities, but they are inherently safe because of a minimum of on-board autonomous code. The cloud or hackers will, for example, not be able to instruct an autonomous car to drive into a wall. Programming such autonomy requires a few very important lines of code, the exact formulation of which is an ongoing challenge. We aim to learn from animals. What simple rules determine their successful behaviour, individually or in a swarm? We have to be realistic of course and accept that some robot accidents will happen, which is no different than society currently accepting accidents with animals and non-autonomous cars. The acceptance of autonomous robots is a challenge on its own as we humans have grown unaccustomed to such autonomy. People need to be re-educated. For this purpose, we collaborate with the faculty of Industrial Design. A first step may be the implementation of haptic feedback. Think of a force-feedback throttle that discourages tailgating, unless your cloud data indicates it to be safe. In ten years’ time we haven’t only grown accustomed to autonomy, we will demand it. Robots will have been accepted by society as a new and tamed animal species. This requires robots to demonstrate animal behaviour. Delfly and ZEBRO, two robots developed at TU Delft, move predictably because they appear to have some kind of head indicating their direction of motion. A quadcopter on the other hand is unpredictable and frightening as it lacks this characteristic. In 2050 the role of robots is one of assistance. Think of a team consisting of a healthcare professional and his or her robot. Perhaps they even received training and education together. The robot will do the menial tasks increasing the opportunities for the professional to express empathy. It has to be a human providing warmth and empathy, although I have to admit that it wasn’t easy when I parted with my ten year old Skoda 130. It wasn’t even a robot. Text: Martijn Engelsman | Photo: Mark Prins

Solving the puzzle: a small scanner for suspicious moles

Researcher Aleksandar Jovic is working on a pen-shaped device that your doctor can use to scan suspicious moles for skin cancer. It promises to be much cheaper and easier than the current method that can only be used by dermatologists. If you find a suspicious spot on your skin or have a mole that suddenly starts itching or bleeding, you obviously want to have it examined as quickly as possible. In that case, your doctor will often refer you to a hospital or a clinic. There, dermatologists with special equipment called Optical Coherence Tomography (OCT) scanners, will determine whether the moles require further examination. But this equipment is slow and expensive, costs between EUR 50,000 and 100,000 and takes up a lot of space. This is why this equipment can only be found in hospitals. But an alternative method is possible. “We are developing a handy portable device, a miniature version of the hospital equipment, that is also less expensive,” says PhD candidate Aleksandar Jovic from the EEMCS faculty. Reduced healthcare costs Your own doctor can use a device like this to scan moles in his or her own office. “The scan is then sent to a dermatologist, who decides if further investigation is necessary,” says Jovic. This has a number of advantages. It clarifies the situation for the patient more quickly and there is no need to visit the hospital or clinic. It also takes less time for the dermatologist, which reduces healthcare costs. “The aim is that this device should be as easy to use as taking a photo with your camera,” says Jovic. But this is still the future, because the device is currently being developed. However, Jovic recently presented a prototype, a microelectromechanical system (MEMS) silicon chip, several millimetres in size, that can scan the skin. His research is part of a broader trend in remote healthcare known as telemedicine. It involves data being collected from different places and shared with a specialist. The idea for the research came from the Spanish company MedLumics, whose founders include someone who was awarded his PhD at TU Delft. They were convinced that the technology found in the sluggish equipment used by dermatologists could be made within a much smaller device. This became the starting point of Jovic's research, which is being fully-funded by the European Union. Ingenuity In recent years, he has been working closely with the Spanish company to refine the technology. It has not been an easy process. “The large-scale OCT scanner works in a similar way to a paper scanner that many people have at home or at work. Imagine you want to scan a document. You place it on a glass plate. Light goes across the paper and the scanner creates an image. The standard OCT using mirrors and lenses works in the similar way, but with much more precision. This is why they are more expensive,” says Jovic. But how do you make it so small that it fits into a pen? That was the difficult part of Jovic's work. In a scanner, you use a mirror and separate lenses. “But what happens if we combine them? And create a device in which all the components are completely integrated into a small silicon chip? It not only saves space, but also a lot of time and money. That's what we have been working on.” Making a lens and mirror in one is a new approach. Jovic says it was like reinventing the wheel. “Thanks to the potential of the MEMS technology, the silicon chip can do the same as the large-scale equipment. The trick is how a light waveguide, a micro meter-scale mirror and a lens are connected into one single block which is then connected to the rest of the silicon chip. Therefore, they do not have to move relative to each other like in standard OCT scanners. Instead they move all together thus moving the light beam as well.” His research is part of a broader trend in remote healthcare known as telemedicine. It involves data being collected from different places and shared with a specialist. Making a lens and mirror in one is a new approach. Thanks to the potential of the MEMS technology, the silicon chip can do the same as the large-scale equipment. Puzzle Jovic takes the chip out of the cabinet with pride. The chip was completely fabricated in Else Kooi Laboratory at TU Delft. He recently presented it at a conference. “We have not yet tested it on skin. But we have demonstrated that you can use it to scan,” he says. “We do eventually aim to test it on skin, which is definitely possible.” Jovic has made important progress in the development of a pen-shaped device that may soon be available at your doctor's practice. He has demonstrated that you can conduct the scanning process in a machine that fits into your hand. Because the device is so small, the layout design of the MEMS chip was crucial. Which component should be placed in which precise location on the chip and how do you connect it to other components? Like an architect, Jovic was continually developing the design and layout. “In the mornings, I often went straight to my whiteboard to work on the layout. I was continually working to find the smartest way of building and connecting everything to each other. My colleagues helped me a lot. I was able to brainstorm with them and they gave me advices. In the last few years, I have continually been consulting with other researchers and colleagues from TU Delft and MedLumics.” It was one big puzzle to solve, he says. Everything needed to fit perfectly, and now it finally does. He hopes to see the research continue. If there is new funding for the project, an improved prototype should be possible within a year, Jovic believes. It should be able to reach the market within three to five years. “Then doctors will soon be able to use it in their practice and make a rapid diagnosis with this device, simply by holding it above a suspicious mole.” Text: Robert Visscher | Image: Mark Prins

Electrical implants

Since the introduction of the pacemaker in 1958, much has changed in the world of electrical stimulation. Whereas the first electrical implants targeted muscles, the implants of today are flexible and focus mainly on the nerves in our body. The concept, however, remains unchanged: electrical implants give control back to the body. Vasiliki Giagka, Assistant Professor of Bioelectronics at TU Delft, talks about the past, present and future of her field of research. Electrical implants of today look nothing like the first pacemaker of sixty years ago. They are much smaller now and have a much longer battery life, while being much more versatile. Small active implants now help us to relieve the symptoms of Parkinson’s disease, alleviate pain in different parts of the body and treat incontinence. According to Vasiliki Giagka, personalised electronics will soon be part of the treatment plan of patients with rheumatoid arthritis, asthma and diabetes, and patients with spinal cord injury will be able to walk again in the foreseeable future. A pipe dream? Giagka: ‘There is huge potential. I am sure there are a lot of people who will benefit from this research.’ The plasticity of the spinal cord ‘Neurons in our body connect to form networks. The spinal cord therefore functions as a conduit for communication between muscles and the brain. For a person who suffers from spinal cord injury, some of these connections are interrupted. About thirty years ago, scientists discovered that networks can re-organise, and new connections can be formed. This phenomenon is called plasticity. We can encourage this process through physical training. In 2012, researchers succeeded in getting a paraplegic patient to support his own weight and even to take a step through a combination of electrical stimulation, medicines and training. Since then, more results have been published on a regular basis. I do not know whether we are going to be able to treat all paraplegic patients – that depends on lots of factors – but I believe there is huge potential. I am sure there are a lot of people who will benefit from this research.’ Sinals Electronic implants can be used to control three modalities in the body: to induce signals that are interrupted, to record signals coming from elsewhere in the body to feed them back to another system and to block unwanted signals. ‘Consider how our bladder works,’ explains Giagka. ‘It fills up, so we have to empty it. There needs to be a signal that instructs our body to do this. This is a signal we can induce with electronics at a time we want to. But then we want to know precisely when we want to induce this signal. So we need to record information that reveals how full our bladder is. We can then use this information to extract a signal that tells us that our bladder is full and needs to be emptied. Some people suffer from urinary incontinence: their bladder empties without them having instructed it to do so. We can use electronics to block that signal and restore the mechanism. This way we close the loop.’ Weak points Closing such a loop with electrical stimulation can heal patients with a variety of conditions, such as spinal cord injury. Giagka: ‘Here in Delft, we are doing research into electronics that stimulate the spinal cord to restore locomotion in paraplegic patients. They can then learn to walk again in a more coordinated way.’ Giagka was already working on this line of research when she completed her PhD in England. ‘I fabricated a flexible electrical implant to stimulate the spinal cord. I wanted to avoid using wires to and from the electronics, because they cause inflammation around the spinal cord. We are currently working on minimising the size and power consumption.’ Bioelectronic medicines are even smaller than current implantable devices, with all the technology concentrated in a small housing in a three-dimensional cuff-like form, which can be placed around a nerve. About Vasiliki