Talk by David De Roure for Aeolian Network in October 2021 https://www.aeolian-network.net/

1. David De Roure
Emerging Scholarly Practice and
Scholarly Primitives: a Case Study
in Music and AI

2. Our knowledge infrastructure has evolved to support digital
scholarship, as we adopt new methods which realise the
affordances of the digital – including computation, but also
socio-technical engagement at scale.
The digital musicology community has been “an early adopter”,
establishing new research practices which illuminate possible
futures in other fields.
Today we see the music community increasingly adopting
artificial intelligence techniques in both analysis and
composition, and a growing symbiosis of human and machine.
The story of the talk

3. Edwards,
P.
N.,
et
al.
(2013)
Knowledge
Infrastructures:
Intellectual
Frameworks
and
Research
Challenges.
Ann
Arbor:
Deep
Blue.
http://hdl.handle.net/2027.42/97552
https://jakefagan.myportfolio.com/knowledge-infrastructures-sketch-development
Knowledge Infrastructures:
Intellectual Frameworks and Research Challenges

4. Social Machines
“Real life is and must be full of all kinds of social
constraint – the very processes from which society
arises. Computers can help if we use them to create
abstract social machines on the Web: processes in
which the people do the creative work and the
machine does the administration... The stage is set
for an evolutionary growth of new social engines.
The ability to create new forms of social process
would be given to the world at large, and
development would be rapid.”
Berners-Lee, Weaving the Web, 1999 (pp. 172–175)

5. Crowdsourcing:
Social Machines (meet AI)

6. Example: What’s the Score in the Bodleian?

7. https://web.archive.org/web/20070716071553/http://www.galaxyzoo.org/

8. At long last the ‘Peas’ have been submitted to MNRAS (The
Monthly Notices of the Royal Astronomical Society,). The
‘Peas’ were discovered by users right here in Galaxy
Zoo who noticed a strange class of small green galaxies at
redshifts near z=0.2. A dedicated group of volunteered
collected a sample of these galaxies. Then Kevin
Schawinski found an astronomer (Carie [me :)]) to pull
them together and look at them in detail.
https://blog.galaxyzoo.org/2009/04/14/the-first-volunteer-inspired-galaxy-zoo-paper-is-submitted/

9. Scientists
Talk
Forum
Image
Classification
data reduction
Citizen Scientists
Galaxy
Zoo
launched
in
July
2007
and
engaged
165,000
volunteers

12. In an effort to speed up classifications to cope with the large number of galaxies we
expect to receive from new surveys, we've been working on ways to combine your
classifications with those of machines, inspired by the idea that the combination of both
automatic and human classification may be more powerful than either alone. If you
choose the 'Enhanced' work flow, you will be much more likely to see the top 100 galaxies
our galaxy-classifying robot thinks it needs help with in order to improve. All galaxies will
be seen by at least a few volunteers to make sure we aren't missing anything. If you'd
rather just see a random selection of available galaxies, choose 'Classic’.

13. Community
Software
Supercomputer
Digital Music
Collections
Student-sourced
ground truth
Community
Software
Linked Data
Repositories
Supercomputer
23,000 hours of
recorded music Music Information
Retrieval Community
SALAMI
JCDL 2013
J.
Stephen
Downie
Ichiro
fujinaga

15. https://britishlibrary.pubpub.org/the-collective-wisdom-handbook-perspectives-on-crowdsourcing-in-cultural-heritage---community-review-version

16. AI-Sourcing:
Human replacement vs co-creation

17. https://www.jstor.org/stable/1513376
https://ess.org/ness/2021/1/21/caa-release-george-e-lewis-voyager

18. "Even if you don’t think machines can be creative
by themselves, they can potentially be creative
friends… You can imagine a situation when you’re
having a conversation with a machine offering
prompts as a critical, creative accomplice.”
Mark d’Inverno
Vortex
Jazz
Club,
October
2016
“The first concert consisting almost entirely of
music composed by artificial intelligence"
Geraint Wiggins

19. Barbican,
9
Mar
2019
The Eternal Golden Braid
RNCM,
13
June
2019
The Voice of the Machine
Royal
Institution
Christmas
Lectures
2019
Secret & Lies
The Hidden Power of Maths
Marcus du Sautoy Robert Laidlow
Lecture 3 How can we all win? (at 57mins)

20. Three Entistatios - BBC Philharmonic Orchestra
(Robert Laidlow, 2019)
https://www.youtube.com/watch?v=839rF8pucmY

21. “The first movement is an exploration of the beginning of the
learning process, where an AI has not yet learned what we
consider the basics of coherent composition and musical technique.
The second movement is a conversation between myself, the
composer, and a state-of-the-art AI network that is fully trained
upon my music. I would write music, then the AI would respond
in a style it had learned from studying my previous works. I
decided which of its ideas were interesting, reworked them into
new forms, and then gave this as the algorithm’s next input.
The final movement moves more towards science fiction. It is an
imagined soundscape of a future-AI that has reached ‘singularity’.

22. Milton Court Concert Hall, 2 Nov 2019
Alter | PRiSM led by Robert Laidlow
Alter was premiered by the Britten Sinfonia and Marta Fontanals-
Simmons, conducted by William Cole, at the Barbican.
https://www.youtube.com/watch?v=L1mQGaNmfUM

23. Ada Lovelace
Charles Babbage
Bodleian Libraries
19th Century AI
21st Century AI
Audience
Human composer
Robert Laidlow
Alter Text – PRiSM team led by Robert Laidlow
Commissioned by Barbican, performed by Britten Sinfonia with
Marta Fontanals-Simmons, conductor William Cole, November 2019.
Mezzo-soprano
Marta Fontanals-Simmons
prompts
Images © Clay Mathematics Institute

24. Robert Laidlow
Hybrid Chain model

25. Our scholarly workflows now engage the crowd and
the AIs, with increasing ease.
With both crowdsourcing and AI, we are
outsourcing work (typically in order to scale).
What are the hidden costs of these approaches?
Crowdsourcing vs AI-sourcing

26. An exercise for the reader: what are the scholarly
primitives for crowd-sourcing and AI?
Scholarly Primitives refer to some basic functions common to
scholarly activity across disciplines, over time, and independent
of theoretical orientation.
These ‘self-understood’ functions form the basis for higher-level
scholarly projects, arguments, statements, interpretations—in
terms of our original, mathematical/philosophical analogy,
axioms.
John Unsworth

27. Thus, it has come to pass not only that
improvisations by creative musical machines
are often indistinguishable from those created
by humans, but also that they need not be so
distinguished.
De
Assis,
P.,
&
Giudici,
P.
(Eds.).
(2021).
Machinic
Assemblages
of
Desire:
Deleuze
and
Artistic
Research
3.
Leuven
(Belgium):
Leuven
University
Press.
doi:10.2307/j.ctv1595mb9
George E. Lewis, 2021

28. International Summer
Courses in Darmstadt
August 2021
http://milker.org/text-score-dataset

29. Thanks to:
Samantha Blickhan, Emily Howard, Robert
Laidlow, Chris Lintott, George E. Lewis, Grant
Miller, Marcus du Sautoy, Victoria Van Hyning,
Pip Willcox
Supported by:
The Alan Turing Institute ’Data Science of Music’
RNCM Centre for Practice & Research in Science &
Music (PRiSM), funded by Research England
Expanding Excellence in England (E3)
Theory and Practice of Social Machines (SOCIAM)
funded by EPSRC