Sony Computer Science Laboratory Paris 15th Anniversary: How computers empower individuals to be creative?, 30th of September 2011

#cslsony on Twitter, identi.ca, Live Cast, TwitterStreamGraphs, Collecta, flickr

Own objectives

  1. prepare CreativiteArtificielleParadigmShift2011
  2. check Sony CSL work

Attended activities

Live Coding performance

Creative Evolutionary Ecosystems

John McCormack (@jonmcc), Monash University, Australia

  • overall much more visual and less technical than his April 2011 project http://diotima.infotech.monash.edu.au/~jonmc/sa/research/creative-ecosystems/
  • history of computer science changed, initially from a military background it gradually democratized and included creative usages
  • why is increasing creativity important?
  • computer can be tools to augment imagination, not just tools to facilitate the production of results or productions
  • evolution is an exploration of the morphological space
    • evolution is an algorithm
    • mention of Richard Dawkins software biomorph from his book the Blind Watchmaker
    • author presentation of his work
      • Evolve 1.2 (1991)
      • Turbulence (1994)
        • some scenes requiring ~1y of computing then to be processed, done in on a supercomputer instead
    • presentation of Eden (2000), e.g. video
      • based on cellular automata
      • including the behavior of spectators as a signal, then having a sense to rely on
      • mostly Lamarckian rather than Darwinian view
  • a lot of this work and similar work is about nature, a nature that ironically we are gradually destroying
  • reverse-engineering process to go from an existing plant, e.g. here a tree, to its model
  • why do we make art, why do other animals make art?
  • basically the evolutionary process is powerful and should be leveraged
  • concluding on Rumi's quote (1273)
  • own question asked
    • can he get a feel of how bounded or not a "creature" in e.g. Eden can be, via the number of parameters or other ways (having in mind also WithoutNotesAugust11#Evolvability)
      • seems not really, it's mostly a coupling of trying and adjusting

Work discovered in WithoutNotesJanuary11#JonMcCormack and added to LinkOfTheMonth#January2011. Back to the Menu

Designing the Future Relationship of Human and Artifacts

Shunji Yamanaka, Keio University, Japan

  • important inspiration from manga
    • working at the intersection of manga and engineering
  • works
  • show of a skeleton automata of an archer made solely of wood and strings
  • interested in lifelikeness
  • own question asked
    • are some mathemical objects, e.g. the golden ratio or Pi, influence, contraint or help is work
      • seemed that it was necessarily easy to match those and that the brain might not have an affinity or an ability to build with those effortlessly
  • note that fuRo:トップ, the Future Robotics Technology Center is working with Leading Edge Design
  • see also the recently started DesignInspiredInnovation and the more general Design

Back to the Menu

Capturing the Essence of Musicians

François Pachet, Sony Computer Science Laboratory Paris and Jeff Szuda, Musician

  • why is jazz in particular interesting for such studies
    • unique style of each player
    • display of individuality
    • having to compose with constraints
  • answer the problem of how to play with musicians one would like to play without player over a fixed tune
    • few details on the tools used, ML, AI, Markov processes, ...
  • demo with bass, drumer and virtuoso, first one by one then all together
  • the real benefit be to always be able to challenge ourselves via the surprises we experience
  • note that he might also be the author of a jazz composer software I tried a long time ago
  • one can wonder if a solo will get a spontaneous applaud from a crowd who knows it is a piece of software

Back to the Menu

Applicable Creativity

Tetsu Natsume and Yoko Honjo, TPO (Technology Promotion Office), Sony Computer Science Laboratories, Inc., Tokyo

  • problem of the death valley between research and industrial production
  • Francois Pachet mentioned his own experience with machine learning and how most researchers focusing on a specific path which yield smaller and smaller returns of accuracy lead him to try other path
    • basically one can wonder if this could be a way to automatize paradigm shift
      • if the seminal metrics of the field reach an assymptote then challenging some of the foundations might be required to go further
  • show of the Tohaku Navi used in Tokyo Museum
    • product name evoking the anime Lain (to confirm)
    • using indoor location sensing through wifi APs (probably triangulation)
    • one can wonder if this could also be used as a learning tool with e.g. Q&A

Back to the Menu

The Painting Fool: Growth

Simon Colton, Imperial College, London

  • directly asking, Can software be creative?
    • points in favors
      • novelty
      • multiple styles
      • judging its own work
      • cultural sense
    • points against
      • no artistic formation received
      • only do what told
      • not task conscious
      • no emotional connection
  • using context-free design grammar, cf context-free grammar discovered before and considered for Seedea:CombinatorialCreativity/GrammarRuleBinary
  • offering a 3D gallery engine, thus able to produce picture of the setup
    • also providing a virtual pedestral
  • demonstration of emotional painting controlled via facial expressions
    • 6 emotions available in the palette
  • demonstration of news extraction then flickr image and several filter
  • showcasing the growth over time with a timeline with productions
  • The Painting Fool - A Computer Artist?
    • productions present in the exhibit

Work discovered in WithoutNotesSeptember11#SimonColton.

  • own question asked
    • in the end this might also be a societal problem thus better understanding the art world social norms might be required, not just computational aesthetics through e.g. existing machine learning projects, actually the history of art in the XXth century might reflect such a trend, incorporating this would somehow take an ontology reproducting the phylogeny perspective, a need to take into account a meta level, to modelize the environment the artist is producing within
      • not tried not sa far, mostly the first part on history of productions, not of art world mechanism
      • pointed to 7 days in the art world

Back to the Menu

Comic Computing - Act Beyond Comic Borders

Hiroaki Tobita, Sony Computer Science Laboratories, Inc., Tokyo

  • impressive demonstration
  • manga as having a specific style and norms
  • creating a tool that reflect those and facilitate interaction
  • show of "cloud computing" as a floating avatar
    • small helium baloon with its video-projector displaying a face
    • made the lab an attractive and curious place

Back to the Menu

Confident Machines

Manfred Hild, Sony Computer Science Laboratory Paris

Back to the Menu

Live Coding performance

  • 3 different and independant instruments
    • solely the synchronization of tiem is shared, still allows to control the speed in one place
    • no hierarchy of somebody producing a beat, somebody else the effects and the last the big picture
  • Scheme Bricks by Dave Griffiths
  • live coding does show the code, not "just" the audio result

Back to the Menu

Overall remarks and conclusions

  • very interesting
  • yet nothing really surprising
  • mostly confirming the knowledge I had on robotics, cognitive science and computer science
    • but more importantly recently of computational creativity

Back to the Menu

See also

Other reviews or coverage

  • here

To do

  1. Sylvain's question on how do all what we see tonight is actually in Sony as a business, or not