Knowledge Change, Failure, Adaptation,
           and Evolution
                                     Jérôme EUZENAT a,1
                        a INRIA and Univ. Grenoble Alpes, France



          Abstract. The wealth of knowledge representations available on the web today un-
          der the form of ontologies will have to evolve because what they represent changes
          and our knowledge about it changes as well. From far away, the terms “change”,
          “adaptation” or “evolution” may seem interchangeable. We discuss their meaning
          under the light of different techniques which have been developed to cope with
          distributed knowledge change: alignment repair, network of ontology revision and
          cultural knowledge evolution. We find that they can be understood through a redef-
          inition of these four terms:
          Change Modification of the condition of use of knowledge through new knowl-
                 edge acquisition (new learning method, new knowledge input, environment
                 change, etc.).
          Failure An unwanted situation raised by the change.
          Adaptation Local operation used to adapt knowledge to observed failure to ac-
                 count for change.
          Evolution Global long term knowledge change due to the repeated application of
                 adaptation operations and selection.
          We show that surveyed techniques can be characterised with respect to differences
          along these dimensions.




1 This is joint work with Fernanda Baiaõ and Kate Revoredo (UNIRIO).