Dr. Simon Colton, Imperial College, UK
Evolving Stuff (A Computational Creativity Perspective)
Computational Creativity is the subfield of AI research where we study how to build software which can take on some of the creative responsibility in arts and science projects. By applying AI techniques in creative domains such as natural language generation, mathematical theory formation, video game design, music composition, the visual arts and graphic design, we aim to push the boundaries of what can be achieved by intelligent software. In recent years, evolutionary programming has emerged as one of the most popular and effective computing paradigms within which to implement systems from a computational creativity perspective. In the talk, I will try and explain the popularity and utility of GA and GP techniques, drawing on experiences I’ve gained from various evolutionary programming projects I’ve been involved with. In the Computational Creativity group at Imperial (www.doc.ic.ac.uk/ccg), we have been evolving stuff for a number of years. In particular, we have undertaken projects to evolve: behaviour trees for controlling the DEFCON video game; painting styles; image filters; design grammars; buildings/cities and transport systems for the Subversion video game; pixel shaders; art-based games; and numerous evolutionary art systems related to The Painting Fool system (software we hope will one day be taken seriously as a creative artist in its own right, as described at www.thepaintingfool.com). Our latest evolutionary art system is called Elvira and will be delivered as an application for the iPad. This has thrown up issues related to the ludic design of EvoArt systems where the user’s aesthetic considerations are utilised as a fitness function. In order to deliver an efficient, playful, environment, we have had to innovate by using pseudo-Lamarckian techniques within the iPad App. We will unveil the first draft of Elvira at the talk. To conclude the talk, I will place evolutionary techniques into the bigger picture of computational creativity research, and I will use the opportunity to describe some fundamental notions which I believe it is important to adhere to when building creative software.
Simon Colton (www.doc.ic.ac.uk/~sgc) is a Reader in Computational Creativity at the Department of Computing of Imperial College, London. He is an AI researcher who studies how to build software which takes on some of the creative responsibility in application domains including video game design, mathematical invention, visual arts and graphic design. His work has spanned various areas of AI research, including theorem proving, machine learning, constraint solving, evolutionary approaches and philosophical aspects. He has published more than 120 papers and his research has been recognised with national and international awards. In the last three years, he has been the principal investigator on six EPSRC-funded projects, which has enabled the building of the Computational Creativity research group at Imperial (www.doc.ic.ac.uk/ccg). Members of the group are involved in projects which aim to: build more sophisticated models of mathematical theory formation; combine reasoning systems so that the whole is more than a sum of the parts; build the next generation of adaptive video games; enable the automatic invention of board games; build an automated artist; add intelligence to browsing-based graphic design; and mine social networking data. Dr. Colton has particular involvement with the video games industry, having worked on funded projects with three different games companies, and by being one of the leaders of the AI and Games Industry/Academia Research Network (www.aigamesnetwork.org).
Prof. Robert John, De Montfort University, UK
Uncertainty Modelling from a Fuzzy Logic Perspective
Fuzzy Logic has a 45 year history of modelling uncertainty, imprecision and vagueness. This talk considers (type-1) fuzzy logic firstly, setting the scene for a discussion about type-2 fuzzy logic. Type-2 fuzzy logic allows for fuzzy membership functions. In this presentation an overview of type-2 fuzzy sets and how the core operations are defined is given. Three applications of type-2 fuzzy sets are described and the talk finishes with a personal perspective on how type-2 fuzzy logic may contribute to future efforts in Computing with Words.
Dr. Joshua Knowles, University of Manchester, UK
Titles and abstracts TBA.
Prof. Richard Olsen, Olsen Ltd, UK.
Road map for high frequency finance: challenges for computational intelligence
The talk starts out by analyzing the causes of the economic crisis and relates these findings to the perceived deficiencies of traditional economic thought. High frequency finance is a new discipline in economics that follows the ubiquitous approach of natural sciences to collect and analyze empirical data and only in a subsequent step discover the underlying explanations. High frequency finance has led to new insights that are revolutionizing the understanding of financial markets and economics at large. There are universally applicable laws that can be used to build financial applications and services for government organizations, financial institutions and investors alike. The talk provides an overview of the task at hand for computational finance.
A short bio:
Richard Olsen is an economic researcher in high frequency finance. He is co-founder and chairman of OANDA, a market maker and information source for currency; is chief executive of Olsen Ltd, an investment manager, and visiting professor at the Centre for Computational Finance and Economic Agents at the University of Essex.