Cellular learning automata is a combination of learning automata and cellular automata. This model is superior to cellular learning automata because of its ability to learn and also is superior to single learning automaton because it is a collection of learning automata which can interact together. In some applications such as image processing, a type of cellular learning automata in which the ...

We consider the problem of stabilizing unstable equilibria by discrete controls (the controls take discrete values at discrete moments of time). We prove that discrete control typically creates a chaotic attractor in the vicinity of an equilibrium. Artificial neural networks with reinforcement learning are known to be able to learn such a control scheme. We consider examples of such systems, di...

Reinforcement Learning (RL) is a machine learning paradigm wherein an artificial agent interacts with environment the purpose of behaviour that maximizes expected cumulative reward it receives from environment. Reward machines (RMs) provide structured, automata-based representation function enables RL to decompose problem into structured subproblems can be efficiently learned via off-policy lea...

Ant colony algorithms are a group of heuristic optimization algorithms that have been inspired by ants foraging for food. In these algorithms there are some agents, the ants, that for finding the suitable solution, search the solution space. On the other hand, Learning Automata is an abstract model that can do finite actions. Each selected action is evaluated by a random environment and the env...

In this paper, a new algorithm which is the result of the combination of cellular learning automata and frog leap algorithm (SFLA) is proposed for optimization in continuous, static environments.At the proposed algorithm, each memeplex of frogs is placed in a cell of cellular learning automata. Learning automata in each cell acts as the brain of memeplex, and will determine the strategy of moti...

Reinforcement learning can be used not only to predict rewards, but also to predict states, i.e. to learn a model of the world's dynamics. Models can be deened at diierent levels of temporal abstraction. Multi-time models are models that focus on predicting what will happen, rather than when a certain event will take place. Based on multi-time models, we can deene abstract actions, which enable...

We demonstrate how to organize 1-dimensional cellular automata into an operad of spaces. The nth term C(k) is the space of radius r = k − 1 automata. The operad composition operation involves both automata composition and shifting of domain. Pointwise operations such as addition of automata become important when we look at the structure of the individual terms in the operad, the spaces of autom...

OF DISSERTATION AN ECHO STATE MODEL OF NON-MARKOVIAN REINFORCEMENT LEARNING There exists a growing need for intelligent, autonomous control strategies that operate in real-world domains. Theoretically the state-action space must exhibit the Markov property in order for reinforcement learning to be applicable. Empirical evidence, however, suggests that reinforcement learning also applies to doma...

Reinforcement learning is a framing of enabling agents to learn from interaction with environments. It has focused generally on Markov decision process (MDP) domains, but a domain may be non-Markovian in the real world. In this paper, we develop a new description of macro-actions for non-Markov decision process (NMDP) domains in reinforcement learning. A macro-action is an action control struct...

The focus of this work is to enumerate the various approaches and algorithms that center around application of reinforcement learning in robotic manipulation tasks. Earlier methods utilized specialized policy representations and human demonstrations to constrict the policy. Such methods worked well with continuous state and policy space of robots but failed to come up with generalized policies....