Reply to “ What are Potential Principles for the Autonomous Development of Value and Motivation ? ”

Welcome to the first issue of the AMD TC Newsletter in 2005. The highlight of this issue is the dialog column guest-moderated by Olaf Sporns: " What are Potential Principles for the Autonomous Development of Value and Motivation? " I would like to thank Olaf and all contributing authors of this column for their insights. The AMD TC Newsletter welcomes more contributions from the AMD TC on research topics interesting to the community. A resolution about changing the name of the ICDL Advisory Board to the ICDL Governing Board (GB) and a procedure to elect the ICDL GB were passed. Each AMD task force presented its past work and future plan. · February, 2005: With Jay McClelland as the ICDL GB Interim Chair and Jochen Triesch as the communication co-Chair, the election of two new members of ICDL GB and the Chair of ICDL GB has taken place. Hideki Kozima and Linda Smith were elected to the board by the membership. are continuing from the ICDL Interim GB for continuity and will rotate off in 1-2 years. John Weng was elected to serve as the first ICDL GB Chair, replacing Jay McClelland, who served as Interim Chair while the election of new members was conducted. Jochen Triesch will continue to serve as communications co-Chair of the GB, serving a three-year term which began in the fall of 2004. One of the goals of research in autonomous mental development is to formulate principles that enable adaptive and generalizable behavior in humans and robots. Clearly, internal evaluations and motivations, in the context of environmental stimuli, action, goals, and plans, play important roles in the development of behavior. As we attempt to build systems that exhibit autonomous development, what have we learned about the role of value and motivation in behavioral development? brain or neural network. Value can be seen as imposing biases on the outcome of interactions between a behaving system and its environment, thus providing a way to evaluate recent behavior and perception. Value systems, as part of a robot's control architecture or an organism's brain, then exert effects that change the propensity or likelihood of such behaviors to occur in the future, in a manner similar to reinforcement learning. Signals produced by value systems represent global evaluations that are internally derived (not externally supplied), and are communicated to widespread regions of the Whether that agenda is derived from drives built …