Fabien Benetou's PIM | ReadingNotes / CreativityAndArtificialIntelligence

Creativity and Artificial Intelligence - A Conceptual Blending Approach by Francisco Camara Pereira - ISBN 3110186098 - De Gruyter 2007

Motivation

Preparing for CreativiteArtificielleParadigmShift2011 and recently discovered it was also used in WithoutNotesSeptember11#OpenCogMonashTalkJuly2011.

Pre-reading model

List the main decisions that should be impacted by reading the book in order to estimate the Seedea:Content/Newconcepts#InformationActionRatio and facilitate rule design as suggested after finishing TheTinkerersAccomplice.

Draw a schema (using PmGraphViz or another solution) of the situation of the area in the studied domain before having read the book.

Reading

1 Introduction
- opening with a quote from Hadamard then Poincare, cf Hadamard
2 Creativity
- overall importance of the work of Wikipedia:Margaret Boden and Wikipedia:Mihaly Csikszentmihalyi
- mention of Sternberg and Lubart's investment theory "creative people are ones who are willing and able to �buy low and sell high� in the realm of ideas." (p27)
- figure 3 (p37) also used Venn diagrams to distinguish different sets of acceptable or not solutions based on Ritchie's criteria
  - as proposed before for CreativiteArtificielleParadigmShift2011
- table 1 (p43) and table 2 (p44) for an overview of creative systems
3 Working with concepts
- clear schema to illustrate Fauconnier and Turner concept blending idea
- clarification of distinction between analogy and metaphors
- analysis of several potential implementation e.g. SME and Sapper
- mention of Hofsdadter, cf IAmAStrangeLoop event though the topic is different
4 A model of concept invention
- generalist model entitled Creative General Problem Solver
  - to compare with my OwnConcepts#UniversalComposer
- note that the difference between bisociation, blending, and conceptual combination are still not obvious
- also assimilated to search (p99)
  - "isn�t this just search in a (complex) space? What more do we have to offer than any other AI model? The answer is simply yes, it is search. And this is an AI model which, as many other AI models, aims to simulate a specific kind of human behavior that has been rarely approached before. In this case, creativity, more specifically concept invention."
  - cf Kyle E. Jennings later 2010 article Search Strategies and the Creative Process, International Conference on Computational Creativity
5 Divago
- no meta-level reasoning
  - see e.g. BeingNoOne or TheEgoTunnel regarding self-models
  - thus non-practical for self-improvement goal of AI
- done in Prolog, impossible to find the source
- using The Generalized Upper Model
- on frames see TheSocietyOfMind
  - could useful frames be generated out of this very wiki?
    - consider the multiple links possible within pages of a group, words in a page, words in a sentence, co-edited pages, ...
- how the mapping m given by φ, i.e. The Mapper, makes meaningful relations is not obvious
  - see http://student.dei.uc.pt/~racosta/jmapper described by the author as an improvement
  - seems be look for proximity of items within a frame
- mention that genetic algorithm in the reasoning module, i.e. The Factory, could be replaced by simulated annealing (p126)
  - as considered for Seedea:Seedea/BackEnd
  - see also CreativiteArtificielleParadigmShift2011#ComputationalComplexityOfCreativity
    - even though P vs NP might not be straight forwardly assimilated to creativity for most, this algorithmic exploration and the choices that have to be made in order to run in realistic time clearly shows a link between computational complexity and creative aims
6 Experiments
- risky mention of CPU speed and result as a measure of comparison with the human system
- critics on the structure required by the blender
  - could structured data e.g. DBPedia help?
7 Conclusion
- theoretical foundations
  - mainly Gilles Fauconnier and Mark Turner Wikipedia:Conceptual blending#Computational_Models
- limits of the implementation
- importance of evaluation
  - mainly Graeme D. Ritchie 2001 Assessing Creativity

Overall remarks and questions

could this technique be applied to imagine what two researchers who have never meet, and eventually could never meet, would produced?
- inspired by reading Avi Wigderson's Knowledge, Creativity and P vs NP after watching Scott Aaronson's Bhul Lecture 2011
  - the later is actually acknowledged by the former in this article
if creativity is mostly about recombination the most important skills might then be efficient re-use and comparison thus rather the ability to connect and incorporate to existing tools or body of knowledge

Synthesis

So in the end, it was about X and was based on Y.

Critics

Point A, B and C are debatable because of e, f and j.

Vocabulary

(:new_vocabulary_start:) new_word (:new_vocabulary_end:)

Post-reading model

Draw a schema (using PmGraphViz or another solution) of the situation of the area in the studied domain after having read the book. Link it to the pre-reading model and align the two to help easy comparison.

Creativity And Artificial Intelligence {ReadingNotes}