Experimental State Splitting for Transfer Learning
نویسندگان
چکیده
Jean is a model of early cognitive development based loosely on Piaget’s theory of sensori-motor and pre-operational thought (Piaget, 1954). Like an infant, Jean repeatedly executes schemas, gradually extending its schemas to accommodate new experiences. We model this process of accommodation with the Experimental State Splitting algorithm. We present the algorithm and demonstrate, in three transfer learning experiments, Jean’s ability to transfer learned schemas to new situations in a real time strategy military simulator.
منابع مشابه
Hierarchical Functional Concepts for Knowledge Transfer among Reinforcement Learning Agents
This article introduces the notions of functional space and concept as a way of knowledge representation and abstraction for Reinforcement Learning agents. These definitions are used as a tool of knowledge transfer among agents. The agents are assumed to be heterogeneous; they have different state spaces but share a same dynamic, reward and action space. In other words, the agents are assumed t...
متن کاملTransfer Learning Using Experimental State Splitting and Image Schemas
Jean is a model of early cognitive development based loosely on Piaget’s theory of sensori-motor and preoperational thought (Piaget 1954). Like an infant, Jean repeatedly executes schemas, gradually extending its schemas to accommodate new experiences. Jean’s environment is a simulated “playpen” in which Jean and other objects move about and interact. Jean’s cognitive development depends on sev...
متن کاملLearning and Transferring Action Schemas
Jean is a model of early cognitive development based loosely on Piaget’s theory of sensori-motor and pre-operational thought. Like an infant, Jean repeatedly executes schemas, gradually transferring them to new situations and extending them as necessary to accommodate new experiences. We model this process of accommodation with the Experimental State Splitting (ESS) algorithm. ESS learns elemen...
متن کاملUtilizing Generalized Learning Automata for Finding Optimal Policies in MMDPs
Multi agent Markov decision processes (MMDPs), as the generalization of Markov decision processes to the multi agent case, have long been used for modeling multi agent system and are used as a suitable framework for Multi agent Reinforcement Learning. In this paper, a generalized learning automata based algorithm for finding optimal policies in MMDP is proposed. In the proposed algorithm, MMDP ...
متن کاملPerformance Analysis of cooperative SWIPT System: Intelligent Reflecting Surface versus Decode-and-Forward
In this paper, we explore the impacts of utilizing intelligent reflecting surfaces (IRS) in a power-splitting based simultaneous wireless information and power transfer (PS-SWIPT) system and compare its performance with the traditional decode and forward relaying system. To analyze a more practical system, it is also assumed that the receiving nodes are subject to decoding cost, and they are on...
متن کامل