Evolving Neural Networks to cus
نویسنده
چکیده
Neural networks were evolved through genetic algorithms to focus minimax search in the game of Othello. At each level of the search tree, the focus networks decide which moves are promising enough to be explored further. The networks effectively hide problem states from minimax based on the knowledge they have evolved about the limitations of minimax and the evaluation function. Focus networks were encoded in markerbased chromosomes and were evolved against a full-width minimax opponent that used the same evaluation function. The networks were able to guide the search away from poor information, resulting in stronger play while examining fewer states. When evolved with a highly sophisticated evaluation function of the Bill program, the system was able to match Bill’s performance while only searching a subset of the moves.
منابع مشابه
Chronic Stress Alters Behavior in the Forced Swim Test and Underlying Neural Activity in Animals Exposed to Alcohol Prenatally: Sex- and Time-Dependent Effects
Dysregulation of the hypothalamic-pituitary-adrenal (HPA) stress response has been suggested to play a role in vulnerability to stress-related disorders, such as depression. Prenatal alcohol exposure (PAE) may result in HPA dysregulation, which in turn may predispose individuals to the effects of stress exposure throughout life, and increase their risk of developing depression compared to unexp...
متن کاملEvolving Fuzzy Neural Networks - Algorithms, Applications and Biological Motivation
In the paper, the ECOS (Evolving Connectionist Systems) framework is used to develop a particular type of evolving neural networks evolving fuzzy neural networks EFuNNs. They can be trained in an on-line, incremental mode and are several order of magnitude faster than the ordinary neural networks trained with the backpropagation algorithm. This is illustrated on the task of adaptive phoneme rec...
متن کاملDynamic Channel Units Allocation in Td- Cdma/tdd System
Hierarchical Genetic Algorithms (HGA) as a tool for search and optimizing methodology have now reached a mature stage. The minimum resource facility to carry user traffic, termed a channel unit (CU), is composed of a one time-slot and one code in the TD-CDMA/TDD system. The control of the number of CUs depends on the traffic load solves varied and asymmetrical traffic problems in the 3G system....
متن کاملAdaptive Neural Networks, Gene Networks, and Evolutionary Systems – Real and Artificial Evolving Intelligence
The paper presents an integrated approach to building evolving artificial intelligent systems in terms of evolving connectionist systems (ECOS) that capture principles from neural networks, gene interaction networks and evolutionary systems. The ECOS can be used to solve complex problems from computational biology that is illustrated on a simplified gene regulatory network modeling problem. The...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1999