More on learning
When asking Eva Jablonka from collaboration, she contacted me with Mikhail Burtzev, who seem sto be a bright young scientist specialized in the application of agent-based computer simulations to uncover principles ruling the evolution of the organization of social systems. The model system as agents living in a defined world (following the Artificial Life approach). he is interested in the following topics:
- Warfare strategies in pre-state societies.
- How the private property was born in evolutionary terms.
- Evolution of cooperative strategies.
- The role of learning in evolution.
The learning device of the agent is a neural network and the genoma is coded in terms of the weights of the neural network, which is to say Mikhail Burtzev follows the approach followed by the tenants of evolutionary computing. The must interesting thing is his research is what he does top uncover evolutionary principles underlying the evolution of cooperative behavior.
I want to know if his interest in evolutionary strategies concerning predating strategies similar to the ones displayed by wolfs has produced new results. I think that some of the approaches followed by Mikhail can be applied to evolutionary robotics problems. I think that the best thing to do is to invite Mikhail to go to Mexico and to give us short course in what he is doing.
What I need to know about all these things concerning learning in evolution
I have several possibilities concerning these topics. In order to evaluate learning in evolution, a possibility will to follow:
- A philogenetyic approach (i.e. to track through bio-informatic approaches the way learning has been replaced by genetic adquisition), but that is not easy because all we know when using this approach come from the genetic system and the memory of fast learning is lost in this information). It is not easy to see this question following this approach.
- Working with computer simulations to see how evolution is changed by learning, i.e. following the Burtzev approach.
- Working with real robots, following an evolutionary robotics approach. the problem in thsi case is that learning would be reduced to neural-systems.
- Working with artificial evolution of uni-cellular organism, i.e. bacteria, or with more complex organism like yeast. In this case the question is how to translate the Burtzev approach to an experimental model to measure the way the organism acquire new components to confront in a better way the environmental changes. I need to speak with Juan Pedraza about these things. i will see him next week (this could be a good approach to work with Alex de Luna).
- Work with an intermediary system, i.e. social system and track the way learning changes the social structure. A possibility is to track the way civilization is conceived as a evolutionary platform to improve adaptation through learning. That is not a bad idea. I need then to continue working with the papers by Jessica Flack on social niches acquisition. This is very important
Well, it is tiem to pass to the definition of the project to be presented to Julio Frenk and To Esther Orozco.
The project around medical questions
I can imagine that it would be nice to have an Institution working in the following questions:
- Cancer dynamics (cellular behavior associated to the application of specific drugs), based on the manipulation of signaling pathway associated to cancer states.
- Synthetic biology approaches to uncover cancer dynamics (I must not forget to read the papers by Sabrina Spencer and to consider it in my proposal).
- A possible aproach would be to have somthing similar to the agreement that Telmex have with MIT, but in this case with Harvard, working around my topics of interest. Which is to say, to have Harvard as the institution to develop the research, working with Mexican posdocs and PhD students, and spreading the acquired knowledge with medical doctors in Mexico (I think that this is the right thing to do). In this propject pablo coulk contribute with something about chimertic proteins as drugs to be applied against cancer.
- Include something about neuro-degenerative diseases and protein folding.
To propose something to Esther Orozco
In this case an effort must be focused around MIT (Media Lab) interests concerning the application of technologies oriented to mobile computing system to:
- Track the way social projects work with the people (creation of efficient databases). It would be nice to have something like an agreement between the Media lab and the Cinvestav, financed by the Instituto de Cienci ay Tecnolog'ia del Distrito Federal (this would be very smart, since coulbe give a very important projection to the activities of the institute in Mexico and will allow me to have ressources for important projects).
- Provide solutions to improve the quality of life of poor people (the micro-finance system seems to be a good idea).
- Do not forget the idea about the product space and propose a project to Ricardo Houssman from JKF School of Government, and also consider network approaches to the understanding of social phenomena in Mexico City. It would be then not stupid to contact people from CIDE, ITAM and UNAM, to include in this project. that is not a bad idea. I need then to contact Esther and to act as his representative with the people here at Harvard and at MIT.
All this stuff must be done before christmas day.
|
The Notebook of JCMG
Main project page
Previous entry Next entry
