Carlos León
On 22/February/2016 the musical “Beyond the fence” saw the light in London. It was claimed that “Beyond the Fence” marks a new dawn in the evolution of musical theatre – the first show in the world conceived and substantially crafted by computer!” (https://artstheatrewestend.co.uk/whats-on/beyond-the-fence/ )
“It is modelled on a statistical study of the ‘recipe for success’ in hit musicals... The process began with a predictive, big data analysis of success in musical theatre… to create a set of constraints to which the musical had to conform, to theoretically optimise chances of success”. (http://beyondthefencemusical.com/ )
This is a singular situation and I think it is worth to reflect on it. So, here are some questions:
- What does the community think will be the consequences for the area of this musical? Is the publicity around it good for the area of CC? Is this how we want the society to perceive research on CC? Does the musical reflect the type of work we develop?
- Also, there are questions that I would like to ask to the people that participated in this effort: What are the main scientific, technical or philosophical contributions to the area of CC that emerge from this experience? What have we learned that we did not know before? I hope to get some answers from them ;-)
You are kindly invited to comment on the topic.
ventura
Joe Corneli
good comparison case. The write-ups (press releases) that I've found
are a bit sparing on details. The following link points to the most
comprehensive summary I've found:
http://english.donga.com/List/3/all/26/528255/1
Notice that the rather impressive headline "Novels created by AI"
doesn't entirely match the facts that are presented in the article:
- short stories (less than 4,000 characters)
- stories were created in coordination with humans and the computer
Carlos León
Dan,
I'm sure there are several submissions to ICCC on the topic.
Since the review process hasn't finished yet, it could be a bit early to ask for comments on this. But given that this is an open forum and it reaches more people than ICCC, I think it would be great if people involved in BtF could share some thoughts or specific information, in particular by answering questions raised by other members. Of course, after the review process is over.
Best wishes,
Carlos
--
Carlos León <cl...@ucm.es>
Assistant Professor
Room 420bis
Department of Software Engineering and Artificial Intelligence
Computer Science Faculty
Universidad Complutense de Madrid
28040 Spain
--
You received this message because you are subscribed to the Google Groups "Computational Creativity Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to computational-creativ...@googlegroups.com.
To post to this group, send email to computational-c...@googlegroups.com.
Visit this group at https://groups.google.com/group/computational-creativity-forum.
To view this discussion on the web visit https://groups.google.com/d/msgid/computational-creativity-forum/d74b6811-4a63-4610-8570-ac9ee01957a3%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
pachetcsl
I confirm that they are several paper submissions about the show.
We'll see how the reviewing process evolves, and possibly organize a specific session during ICCC if some of these papers are accepted.
Otherwise, we can have this debate in this forum.
So let's wait until mid-April !
cutg...@gmail.com
There's been some hyperbolic AI reporting lately but this one really was something.
Mark d'Inverno
Mark d'Inverno
I’ve never been to the computational creativity conference before but went to ICCC 2016 in Paris and really enjoyed the whole event. As part if proceedings I was asked to chair a session on what we might have learned from some of our community being involved in the “Beyond the fence” musical. Most people will know what it was but it was a musical that had certain elements generated by technologies from the ICCC community. In many respects this seemed like a great achievement in for the computational creativity field- a west end show running for 2 weeks gaining critical attention from the national press forming the denouement of a 2 part series commissioned and broadcast by Sky Arts entitled “The Computer Says Show” which explored a range of issues around the role of technology in the creative process – what could the computer contribute autonomously? How could human creative work with the latest technologies? What future role could we envisage systems from computational creativity playing a role in professional art productions such as a musical.
I deliberately invited members from a wide range of different perspectives as there is nothing duller than fierce agreement on panel sessions. These included several of the leading academics in computational creativity leading research projects that featured in the development of constituent elements of the musical, Simon Colton and Francois Pachet, Archie Norman - the chief exec of the company Wingspan who had successfully pitched and made the programme for Sky Arts, the historian and psychologist Arthur Still, as well as Melly Still, a leading theatre and opera directors (e.g. National Theatre, Royal Shakespeare Company, Glyndebourne Opera House, Broadway) who had proposed the producer/director team of the musical to Wingspan.
We did record the full panel discussions and I hope ICCC make them available at some point but here are a few of my thoughts about what I learned from my limited involvement with the show and from chairing the panel about opportunities for research in this area. In what way does this research field need to progress in order that we could sustain interest in the idea of artistic content being generated (at least in part) by machine? How would we move from a TV programme exploring the potential and limitation of “creative systems” to generate a musical to the situation where west end producers and creative teams were coming to us directly to use our technologies?
A big step for sure – but an important one for us to consider. Starting not with the question of “I wonder what my system can do” but “how can I help current creative teams in ways that would excite and inspire them?”.
One of the things that was noticeable was that a lot of the systems that produced content were “one shot” calculations. So a system might (for example) generate a dozen or more scores and the creative team would sift through the content and see what was useable. Often the human creatives would take a few bars in order to seed an idea that they could complete into a full written song. In most cases there wasn’t a “creative conversation” between human and machine that one would normally expect between humans when devising a show. And that seems to be what the next step for us should be: how do we build systems where these creative conversations can take place in ways which are meaningfull for human creative?
As I have written elsewhere it still seems we are some way of the stage where content generated autonomously by machine could ever sustain our interest for any period of time. As a computer scientist this is an interesting challenge, but as a musician myself, and as a potential audience member, I find it hard to imagine a scenario where we could sustain interested in solely generated artificial content for very long. The times when something has sustained interest in me is in music performance situations because the human is put under new challenges to work with an autonomous system because it can take them out of their comfort zones and they have to work harder to make things work musically. And for the musical to work in this way they need to embue the system with its own creative agency. They need to give it equal billing to get the best out of themselves and of the unfolding creative collaboration.
So I think that where the future lies is exploring artificial creative agency. This is the idea that machines enable new kinds of creative partnerships for humans. That they stimulate, challenge, provoke us to work in new ways and to produce content that would not have been possible without the system. And, come to that, would not - or could not - have occurred working with any other human collaborator.
So we need to start with the human creative, and build systems that demonstrate this creative agency to creative. Systems which immediately – or at least quickly- open up new opportunities for collaboration where the human creative is happy for the system to take creative control at points in the dialogue. Such systems need agency, and this involves an awareness of the human creative, their goals, their previous content, the way they like to work, the artistic influences of the creative, and also - and this is where it starts to get interesting - influences (algorithmic or human) that could take the human creative into entirely unexplored territories. But I think we need to start with the creative and think about designing systems with the right kinds of agency and flow. Starting with the system and then trying to work out how a human creative might interact with it later seems the wrong way round.
And if we can achieve systems which allow new forms of creative collaboration then I think we would be ready to pitch the next programme for Sky Arts. The programme would then be to understand how human creatives could willingly enter new kinds of dialogues with machines that could help them develop their own creative practice in as yet - for them - unimagined ways. So that the next time a creative team come together to put on a musical, it is they that come to us. They come to us because they recognise they can work in new ways to produce a show that would not have been possible without our systems.