Dr Rahul Savani University of Liverpool, UK
Automated Market Making using Reinforcement Learning
We will describe the design and evaluation of an agent for automated market making in financial markets. The agent was developed using reinforcement learning, a machine learning paradigm of learning through trial and error.
Dr Rahul Savani is a Reader in the Department of Computer Science at the University of Liverpool. He works on problems at the interface of economics/finance and computer science. He has worked extensively on equilibrium computation for game-theoretic models of strategic interaction. He has also worked as a consultant on a variety of algorithmic trading projects. Currently, he leads the Economics and Computation Research Group.
Prof Cipriani Christian The Biorobotics Institute, Italy
Achievements and challenges towards the development of thought controlled prostheses
The hand is the chief instrument of action and sensation in humans. With its elegant anatomical architecture comprising muscles, tendons and articulations it allows us to perform a broad range of tasks ranging from power-tool handling to precise fingertip manipulation. Not only, due to its distributed sensory system counting thousands of modality-specific mechanoreceptors, the hand is an extremely sophisticated instrument which is fundamental for sensing the surrounding environment and for developing fine and adaptive motor control skills. Thus upper limb amputation causes amputees to suffer the loss of both the motor and the sensory function. Technologically advanced “myoelectric” prostheses allow for simple movements to the user (typically one or two active movements or few prehensile patterns) and can be voluntarily operated in a way that is far from being natural. The electrical signals produced by muscle contractions (electromyogram, EMG) are picked-up by surface electrodes placed on the residual limb and processed to activate the movements of the prosthesis, sequentially. These prostheses provide a decent level of grasping function but completely fail when it comes to the sensory function: sensory feedback is not purposely provided to the amputee and hence the prosthesis is felt as a non-corporeal object and worn as a shoe would be worn. As a result of such a gap to the natural limb 30-50% of amputees abandon their prosthesis. In this talk, I will provide an overview of the (ongoing) research carried out by our group at the BioRobotics Institute in this interdisciplinary area. I will present some of the solutions proposed to tackle the fundamental challenges faced by prosthetic designers and engineers: 1) how to achieve physiologic control over multiple degrees of freedom, 2) how to provide intuitive sensory feedback to enhance (closed-loop) control and to promote embodiment of the prosthesis, 3) how to interface with the body to implement this bi-directional information flow and 4) how to build functional artificial limbs suitable for such user-prosthesis interfaces.
Christian Cipriani is a Professor of Bioengineering and Head of the Artificial Hands Area at The BioRobotics Institute, Scuola Superiore Sant’Anna (SSSA), Pisa, Italy. He received the M.Sc. degree in electronic engineering from the University of Pisa, Italy, in 2004 and the Ph.D. in BioRobotics from the IMT Institute for advanced studies, Lucca, Italy in 2008. He has been working at the SSSA since 2005 and –as a Visiting Scientist– at the University of Colorado Denver | Anschutz Medical Campus, in 2012, within many national and international research projects on prosthetics and robotics. His current field of research is (bio)mechatronics applied to the area of upper limb prosthetics. He is interested in mechatronic, controllability and sensory feedback issues of dexterous robotic hands to be used as thought-controlled prostheses. On these topics he has authored 100+ peer reviewed scientific papers, 50+ of which on international journals in the field of prosthetics and rehabilitation robotics, and has coordinated 6 large national/European research projects. He also filed eight patents on related fields and founded a spin-off company of the Scuola Sant’Anna – Prensilia. Dr. Cipriani is the recipient of several awards including a Starting Grant from the European Research Council in 2015, an early career grant by the Italian Ministry of Research in 2011 (FIRB 2010 program), a Fulbright Research Scholar fellowship in 2011 and the d’Auria Award from the Italian Robotics and Automation Association, in 2009.
Prof Erol Gelenbe Imperial College, London
Energy Packet Networks: from Data Transmission to Smart Energy Systems
Energy Packets (EP) are a discrete representation of units of energy that we introduced in 2012. Based on recent work we will show how EPs can be dynamically dispatched in a computer system or network in order to optimise the energy consumption of Clouds and Networks. We will also show how the EP paradigm can be used to represent the energy efficient operation of wireless devices, and discuss how EPs can describe the efficiency of data transmission with quantum devices.
Erol Gelenbe, a Fellow of IEEE and ACM, founded the field of performance evaluation of computer systems and networks in Europe.
A native of Istanbul, Turkey, he is a Fellow of the Science Academies of Belgium, Hungary, Poland, and Turkey, of Academia Europaea, and of the French National Academy of Engineering. He has invented new mathematical models such as G-Networks, and the Random Neural Network, and patented technical inventions such as the Cognitive Packet Network. Awarded Knight of the Legion of Honour (France), and Commander of Merit (Italy), he received the Dennis Gabor Medal of the Hungarian Academy of Sciences, the ACM-SIGMETRICS Life-Time Achievement Award, the Grand Prix France-Telecom of the French Academy of Sciences, and the Oliver Lodge Medal of the IET (London).
Dr. Mark Winands Maastricht University, The Netherlands
The Magic of Monte Carlo Tree Search
Monte-Carlo Tree Search (MCTS) has caused a revolution in computer game-playing the last few years. The most well-known example is the game of Go. MCTS is a best-first search technique that gradually builds up a search tree, guided by Monte-Carlo simulations. In contrast to many classic search techniques, MCTS does not require a heuristic evaluation function that assesses the current board position. In this talk I will discuss its background, basic mechanism, and standard enhancements that have improved the technique considerably. Successful applications of the technique in several domains will be mentioned.
Mark Winands received a Ph.D. degree in Artificial Intelligence from the Department of Computer Science, Maastricht University, Maastricht, The Netherlands, in 2004. Currently, he is an Associate Professor at the Department of Data Science & Knowledge Engineering, Maastricht University. His research interests include heuristic search, machine learning and games. He has written more than eighty scientific publications on Games & AI. Mark serves as an editor-in-chief of the ICGA Journal, associate editor of IEEE Transactions on Computational Intelligence and AI in Games, editor of Game & Puzzle Design. He is a member of the Games Technical Committee (GTC) | IEEE Computational Intelligence Society, and member of working group 14.4 – Entertainment Games | IFIP TC14 on Entertainment Computing.
The Business of Software: Software Asset Management (SAM) and the Software Supply Chain
The business, legal and technical issues related to using software commercially are not for the faint hearted and very often overlooked or ‘postponed’ leading to sometimes disastrous results. In his talk Steve will address the challenges of working in a fast paced product development company with software at its core. The ever changing commercial software business models and entitlement frameworks, open source and intellectual property contamination, security vulnerabilities and the resulting day to day headaches associated with compliance that every enterprise, small and medium business should be prepared for.
Steve Cropper is a seasoned Software business professional with over 30 years in the Software Industry, most of those working in Silicon Valley, CA. Steve spent 15 years working as a Senior Manager at Cisco Systems with responsibilities for Software license procurement, royalty management oversight and Software compliance. Steve is passionate about Software its use, license compliance and its pitfalls.