The department of computing at the Open University has a competitive studentship available for full time PhD study. One of the topics covers stochastic generation of music. Suitable students will have either previous experience of computer music generation in general, or alternatively some knowledge of statistical NLP techniques with some general music experience, such as playing an instrument.
The position is competitive, so please contact me at the earliest opportunity if you wish to apply.
Please feel free to contact me informally for more information, by email or phone.
Dr Robin Laney, Open University, UK
r.c.laneyopen.ac.uk
tel: +44(0)1908 654342
FULL DESCRIPTION
Stochastic Modelling for Computer Music Generation: Bridging the Gap Between Composer and Listener
This project will address the problem of educating a music listener in becoming familiar with new genres of music. The basic idea is to model some works the listener is already familiar with and some from the new genre and to generate works that bridge any cognitive gaps. The proposed approach will apply a scaffolding theory of education [1]. We will build upon existing work in computer music generation.
There exist a number of relatively successful approaches to computer music generation based on the use of stochastic techniques. [2]. One composer who successfully uses a range of computational approaches is David Cope [3] who has published the conceptual details of his approach. This project will look at how some of the various techniques can be evaluated in scientific terms [4] and hence combined and applied in a principled way. Stochastic approaches to music generation use a corpus of existing works to generate new compositions that in some sense fit the corpus. [2,4]. Existing work [2,4] has shown that statistical natural language processing techniques [5] can be applied to music generation. There are however, circumstances, as for natural language processing, where uncertainty becomes problematic.
This project would build on existing approaches by extending them to deal with multiple corpora, and involving the user in uncertainty resolution. The latter, rather than being simply a fix for poor technology, will represent a new context for exploring problem based and inquiry learning [1].
There are further ways to exploit the above ideas. A slight adaptation to the approach would provide for automated music recommendation services [6] in, for example, e-commerce, that go beyond existing approaches. The approach also has value from a musicology point of view, allowing the exploration of how a composers style evolved by selecting corpora from given time-periods and generating a sequence of intermediate works.
References:
[1] Scaffolding and Achievement in Problem-Based and Inquiry Learning: A Response to Kirschner, Sweller, and Clark (2006) Hmelo-Silver, Duncan, & Chinn. (2007). Educational Psychologist, 42(2), 99107
[2] Pearce, M. T. (2005). The Construction and Evaluation of Statistical Models of Melodic Structure in Music Perception and Composition. Doctoral Dissertation, Department of Computing, City University, London, UK.
[3] David Cope (2005). Computer Models of Musical Creativity. Cambridge, MA: MIT Press.
[4] Pearce, M. T. and Wiggins, G. A. (2007). Evaluating cognitive models of musical composition. In A. Cardoso and G. A. Wiggins (Eds.), Proceedings of the 4th International Joint Workshop on Computational Creativity, (pp. 73-80). London: Goldsmiths, University of London.
[5] Manning, C., Schutze, H. (1999) Foundations of Statistical Natural Language Processing. MIT Press.
[6] Li, Q., Kim, B. M., Guan, D. H., and Oh, D. w. 2004. A music recommender based on audio features. In Proceedings of the 27th Annual international ACM SIGIR Conference on Research and Development in information Retrieval (Sheffield, United Kingdom, July 25 - 29, 2004). SIGIR '04. ACM, New York, NY, 532-533.
Supervisors: Robin Laney, Alistair Willis, Paul Garthwaite
Contact Details:
Dr Robin Laney
Senior Lecturer in Computing
Department of Computing,
The Open University,
Milton Keynes, MK7 6AA,
UNITED KINGDOM
Tel: +44 1908 654342
E-mail: r.c.laneyopen.ac.uk
If you want to receive academic resources in your e-mail on daily basis, please subscribe to 10resources-subscribe@yahoogroups.com.
