Formal language

About: Formal language is a research topic. Over the lifetime, 5763 publications have been published within this topic receiving 154114 citations.

Fool's Gold: Extracting Finite State Machines from Recurrent Network Dynamics

Abstract: Several recurrent networks have been proposed as representations for the task of formal language learning. After training a recurrent network recognize a formal language or predict the next symbol of a sequence, the next logical step is to understand the information processing carried out by the network. Some researchers have begun to extracting finite state machines from the internal state trajectories of their recurrent networks. This paper describes how sensitivity to initial conditions and discrete measurements can trick these extraction methods to return illusory finite state descriptions.

Cellular Automata and Language Theory.

Abstract: Cellular automaton A (one-dimensional) cellular automaton is a linear array of cells which are connected to both of their nearest neighbors. The total number of cells in the array is determined by the input data. They are exactly in one of a finite number of states, which is changed according to local rules depending on the current state of a cell itself and the current states of its neighbors. The state changes take place simultaneously at discrete time steps. The input mode for cellular automata is called parallel. One can suppose that all cells fetch their input symbol during a pre-initial step. Iterative array Basically, iterative arrays are cellular automata whose leftmost cell is distinguished. This socalled communication cell is connected to the input supply and fetches the input sequentially. The cells are initially empty, that is, in a special quiescent state. Formal language The data on which the devices operate are strings built from input symbols of a finite set or alphabet. A subset of strings over a given alphabet is a formal language. Signal Signals are used to transmit and encode information in cellular automata. If a cell changes to the state of its neighbor after some k time steps, and if subsequently its neighbors and their neighbors do the same, then the basic signal moves with speed k through the array. With the help of auxiliary signals, rather complex signals can be established. Closure property Closure properties of families of formal languages indicate their robustness under certain operations. A family of formal languages is closed under some operation, if any application of the operation on languages from the family yields again a language from the family. Turing machine A Turing machine is the simplest form of a universal computer. It captures the idea of an effective procedure or algorithm. At any time the machine is in any one of a finite number of states. It is equipped with an infinite tape divided into cells and a read-write head scanning a single cell. Each cell may contain a symbol from a finite set or alphabet. Initially, the finite input is written in successive cells. All other cells are empty. Dependent on a list of instructions, which serve as the program for the machine, the action is determined completely by the current state and the symbol currently scanned by the head. The action comprises the symbol to be written on the current cell, the new state of the machine, and the information of whether the head should move left or right. Decidability A formal problem with two alternatives is decidable, if there is an algorithm or a Turing machine that solves it and halts on all inputs. That is, given an encoding of some instance of the problem, the algorithm or Turing machine returns the correct answer yes or no. The problem is semidecidable, if the algorithm halts on all instances for which the answer is yes.

Formal languages and groups as memory

Abstract: We present an exposition of the theory of finite automata augmented with a multiply-only register storing an element of a given monoid or group. Included are a number of new results of a foundational nature. We illustrate our techniques with a group-theoretic interpretation and proof of a key theorem of Chomsky and Schutzenberger from formal language theory.