Phase-Space Learning
نویسندگان
چکیده
Existing recurrent net learning algorithms are inadequate. We introduce the conceptual framework of viewing recurrent training as matching vector fields of dynamical systems in phase space. Phasespace reconstruction techniques make the hidden states explicit, reducing temporal learning to a feed-forward problem. In short, we propose viewing iterated prediction [LF88] as the best way of training recurrent networks on deterministic signals. Using this framework, we can train multiple trajectories, insure their stability, and design arbitrary dynamical systems.
منابع مشابه
Hybrid Teaching-Learning-Based Optimization and Harmony Search for Optimum Design of Space Trusses
The Teaching-Learning-Based Optimization (TLBO) algorithm is a new meta-heuristic algorithm which recently received more attention in various fields of science. The TLBO algorithm divided into two phases: Teacher phase and student phase; In the first phase a teacher tries to teach the student to improve the class level, then in the second phase, students increase their level by interacting amon...
متن کاملQUANTUM VERSION OF TEACHING-LEARNING-BASED OPTIMIZATION ALGORITHM FOR OPTIMAL DESIGN OF CYCLIC SYMMETRIC STRUCTURES SUBJECT TO FREQUENCY CONSTRAINTS
As a novel strategy, Quantum-behaved particles use uncertainty law and a distinct formulation obtained from solving the time-independent Schrodinger differential equation in the delta-potential-well function to update the solution candidates’ positions. In this case, the local attractors as potential solutions between the best solution and the others are introduced to explore the solution space...
متن کاملThe pathology of Farhangian university internship for present an appropriate model: a mixed study
The purpose of the present research was The pathology of Farhangian university internship for present an appropriate model. Mixed species research was used and Sequential explanatory method. The study population consisted of teachers, teachers' guides and Interns Farhangian university have been Chaharmahal and Bakhtiari Province. Quantitative phase, teachers census method and the other two grou...
متن کاملA Phase Space Approach to Minimax Entropy Learning and the Minutemax Approximations
A.L.Yuille Smith-Kettlewell Inst. San Francisco, CA 94115 There has been much recent work on measuring image statistics and on learning probability distributions on images. We observe that the mapping from images to statistics is many-to-one and show it can be quantified by a phase space factor. This phase space approach throws light on the Minimax Entropy technique for learning Gibbs distribut...
متن کاملManifold Learning-Based phase Space Reconstruction for Financial Time Series
Phase space reconstruction of financial time series has become an effective approach for financial system analysis. Since the phase space obtained by the traditional method had the large embedding dimension and was susceptible to noise, this paper proposed a novel phase space reconstruction approach based on manifold learning, called manifold learning-based phase space reconstruction (MLPSR), w...
متن کاملWave Propagation at the Boundary Surface of Inviscid Fluid Half-Space and Thermoelastic Diffusion Solid Half-Space with Dual-Phase-Lag Models
The present investigation deals with the reflection and transmission phenomenon due to incident plane longitudinal wave at a plane interface between inviscid fluid half-space and a thermoelastic diffusion solid half-space with dual-phase-lag heat transfer (DPLT) and dual-phase-lag diffusion (DPLD) models. The theory of thermoelasticity with dual-phase-lag heat transfer developed by Roychoudhar...
متن کامل