CogTrust è un progetto di ricerca che utilizza JFCM per la parte “intelligente” del loro modello, e rappresenta uno degli esempi di come, con la multi-disciplinarietà e l’utilizzo di strumenti e tecnologie esistenti open source, si possano ottenere grandi risultati, contribuendo al miglioramento degli strumenti stessi.
Leggendo le loro risposte da sviluppatore e “papà” di JFCM mi sono sentito molto orgoglioso, ma soprattutto mi prudono le mani perché mi hanno fatto venire in mente un sacco di idee nuove!
Read more →
Cog-Trust is an interesting project that mixes sociology, artificial intelligence and information technology to provide an environment where different delegation strategies can be tested and analyzed using software agents.
We have planned an interview with the authors, in the mean time don’t miss the project’s website:
Cog-Trust Sourceforge Page
Read more →
Main development Recently I was able to work on a more regular basis on JFCM-core, so I could start fixing the most important bugs. You can see it also from the increased number of SVN commits, mainly on JFCM-core module.
My main goal in the short period is to fix all outstanding bugs so we have a “Generally Available” (GA)** version** that could be used, tested, analyzed, optimized, etc. in the real world, for real problems.
Read more →
I had the time to finally fix the most important bug in JFCM.
In short: calculation of concept outputs, up to version 1.0.3, was incorrect!
Now BaseConceptActivator.activate() uses a two-step algorithm:
recalculates incoming connections’ output; calculates Concept’s output. Go download it!
Read more →