Constrained Second - Order RecurrentNetworks for Finite - State
نویسندگان
چکیده
This paper presents an improved training algorithm for second-order dy-namical recurrent networks applied to the problem of nite-state au-tomata (FSA) induction. Second-order networks allow for a natural encoding of nite-state automata in which each second-order connection weight corresponds to one transition in a nite-state automaton. In practice , however, when trained using gradient descent, these networks seldom assume this type of encoding and sophisticated algorithms must be used to extract the encoded automata. This paper suggests a simple modiica-tion to the standard error function for second-order dynamical recurrent networks which encourages these networks to assume natural FSA en-codings when trained using gradient descent. This obviates the need for cluster-based extraction techniques and provides a simple method for guaranteeing the stability of the network for arbitrarily long sequences. Initial results also suggest that fewer training strings must be presented to achieve convergence using the modiied error.
منابع مشابه
Second order sensitivity analysis for shape optimization of continuum structures
This study focuses on the optimization of the plane structure. Sequential quadratic programming (SQP) will be utilized, which is one of the most efficient methods for solving nonlinearly constrained optimization problems. A new formulation for the second order sensitivity analysis of the two-dimensional finite element will be developed. All the second order required derivatives will be calculat...
متن کاملVariational Calculations for the Relativistic Interacting Fermion System at Finite Temperature: Application to Liquid 3He
In this paper, at first we have formulated the lowest order constrained variational method for the relativistic case of an interacting fermion system at finite temperature. Then we have used this formalism to calculate some thermodynamic properties of liquid in the relativistic regime. The results show that the difference between total energies of relativistic and non-relativistic cases of liqu...
متن کاملThe Exact Solution of Min-Time Optimal Control Problem in Constrained LTI Systems: A State Transition Matrix Approach
In this paper, the min-time optimal control problem is mainly investigated in the linear time invariant (LTI) continuous-time control system with a constrained input. A high order dynamical LTI system is firstly considered for this purpose. Then the Pontryagin principle and some necessary optimality conditions have been simultaneously used to solve the optimal control problem. These optimality ...
متن کاملNonlinear Guidance Law with Finite Time Convergence Considering Control Loop Dynamics
In this paper a new nonlinear guidance law with finite time convergence is proposed. The second order integrated guidance and control loop is formulated considering a first order control loop dynamics. By transforming the state equations to the normal form, a finite time stabilizer feedback linearization technique is proposed to guarantee the finite time convergence of the system states to zero...
متن کاملWeak-duality Based Adaptive Finite Element Methods for Pde-constrained Optimization with Pointwise Gradient State-constraints
Adaptive finite element methods for optimization problems for second order linear elliptic partial differential equations subject to pointwise constraints on the `-norm of the gradient of the state are considered. In a weak duality setting, i.e. without assuming a constraint qualification such as the existence of a Slater point, residual based a posteriori error estimators are derived. To overc...
متن کامل