Browsing by Author "Wang, Jing"

Now showing 1 - 8 of 8

A New Learning Algorithm for a Fully Connected Neuro-Fuzzy Inference System
(2014-10-01)
Chen, C. L. Philip
;
Wang, Jing
;
Wang, Chi-Hsu
;
Chen, Long
;
電機資訊學士班
A traditional neuro-fuzzy system is transformed into an equivalent fully connected three layer neural network (NN), namely, the fully connected neuro-fuzzy inference systems (F-CONFIS). The F-CONFIS differs from traditional NNs by its dependent and repeated weights between input and hidden layers and can be considered as the variation of a kind of multilayer NN. Therefore, an efficient learning algorithm for the F-CONFIS to cope these repeated weights is derived. Furthermore, a dynamic learning rate is proposed for neuro-fuzzy systems via F-CONFIS where both premise (hidden) and consequent portions are considered. Several simulation results indicate that the proposed approach achieves much better accuracy and fast convergence.
Finding the Capacity of Fuzzy Neural Networks (FNNs) via Its Equivalent Fully Connected Neural Networks (FFNNs)
(2011)
Wang, Jing
;
Wang, Chi-Hsu
;
Chen, C. L. Philip
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
The capacity of Fuzzy Neural Network (FNN) is explored in this paper. The FNN is first transformed into an equivalent fully connected three layer neural network, or FFNN, via a new approach proposed in this paper. Then the lower and upper bounds of FNN can be found. To check the validity of the theoretical bounds, an example is illustrated with the trainings to yield excellent capacity matching with the theoretical bounds. This new finding has its emerging values in all engineering applications using FNN, such as intelligent adaptive control, pattern recognition, and signal processing, ... , etc.
Improved efficiency of blue phosphorescence organic light-emitting diodes with irregular stepwise-doping emitting layers
(2013-03-01)
Liu, Jun
;
Wang, Jing
;
Huang, Saijun
;
Chen, Hsi-An
;
He, Gufeng
We demonstrate highly efficient blue phosphorescence organic light-emitting diodes (PHOLED) with irregular stepwise-doping emitting layers (EMLs). High efficiency is realized by inserting a high-doping EML between two relatively low-doping EMLs. Peak luminous efficiency and power efficiency of 19.2cdA1 and 18.0lmW1 are achieved, while those of a continuous stepwise-doping device are 9.8cdA1 and 8.6lmW1. The dramatic efficiency enhancement is mainly attributed to the high holeelectron recombination probability and suppressed triplet excitons quenching by fine tuning the doping concentrations in the EMLs.
On the BP Training Algorithm of Fuzzy Neural Networks (FNNs) via Its Equivalent Fully Connected Neural Networks (FFNNs)
(2011)
Wang, Jing
;
Wang, Chi-Hsu
;
Chen, C. L. Philip
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper, Fuzzy Neural Network (FNN) is first transformed into an equivalent fully connected three layer neural network, or FFNN. Based on the FFNN, BP training algorithm is derived to tune both the premise and consequent part of FNN. Illustrative examples are presented to check the validity of the proposed theory and algorithms. Simulation achieves satisfactory results. Developing BP training algorithm for FNN via its equivalent FFNN has its emerging values in all engineering applications using FNN, such as intelligent adaptive control, pattern recognition, and signal processing..., etc
On the Conjugate Gradients (CG) Training Algorithm of Fuzzy Neural Networks (FNNs) via Its Equivalent Fully Connected Neural Networks (FFNNs)
(2012-01-01)
Wang, Jing
;
Chen, C. L. Philip
;
Wang, Chi-Hsu
;
電機工程學系
;
Department of Electrical and Computer Engineering
In this paper, Fuzzy Neural Network (FNN) is transformed into an equivalent fully connected three layer neural network, or FFNN. Based on the FFNN, conjugate gradients (CG) training algorithm is derived to tune both the premise and consequent part of FNN, and apparently increase the speed of convergence. Illustrative examples are presented to check the validity of the proposed theory and algorithms. Simulation achieves satisfactory results. Developing CG training algorithm for FNN via its equivalent FFNN has its emerging values in all engineering applications using FNN, such as intelligent adaptive control, pattern recognition, and signal processing., etc
On the Special Issue for the 12th World Filtration Congress
(2017-11-01)
Pui, David Y. H.
;
Hoeflinger, Wilhelm
;
Wang, Jing
;
Tsai, Chuen-Jinn
;
環境工程研究所
Polarization-Insensitive Remote Access Unit for Radio-Over-Fiber Mobile Fronthaul System by Reusing Polarization Orthogonal Light Waves
(2016-02-01)
Wu, Beilei
;
Sung, Jiun-Yu
;
Yan, Jhih-Heng
;
Xu, Mu
;
Wang, Jing
To reduce the complexity and cost of remote access units (RAUs), colorless transport architecture is preferred in a wavelength-division multiplexing (WDM) system. The wavelength-reuse technique is commonly adopted to realize colorless RAUs. However, in traditional approaches, RAUs employ crystal-based Mach-Zehnder modulators (MZMs) to modulate the distributed carriers. Hence, polarization tracking systems are necessary to align the polarization state of the distributed carriers with the MZM principal axis. This increases the cost of the RAUs. In this paper, a full-duplex millimeter-wave (MMW) radio-over-fiber (RoF) access architecture employing polarization-insensitive RAUs is proposed. In our proposed RAUs, the upstream signals are generated using polarization-orthogonal lights, and the performance of the upstream signals is independent of the polarization state of the MZM input lights. The proposed system is mathematically formulated and experimentally demonstrated. The 1.2-Gb/s quadrature phase-shift keying (QPSK) orthogonal frequency-division multiplexing (OFDM) downstream signal at the 60-GHz band and 5-Gb/s on-off keying (OOK) polarization-insensitive upstream signal over a 15-km standard single-mode fiber (SSMF) are demonstrated. In our results, a polarization insensitive RAU is successfully obtained. The power penalty resulting from the interference of downstream signal is about 0.4 and 0.6 dB with a bit error rate (BER) of 10(-9) in back-to-back and 15-km fiber transmission, respectively.
The Bounded Capacity of Fuzzy Neural Networks (FNNs) Via a New Fully Connected Neural Fuzzy Inference System (F-CONFIS) With Its Applications
(2014-12-01)
Wang, Jing
;
Wang, Chi-Hsu
;
Chen, C. L. Philip
;
電機資訊學士班
;
Undergraduate Honors Program of Electrical Engineering and Computer Science
In this paper, a fuzzy neural network (FNN) is transformed into an equivalent three-layer fully connected neural inference system (F-CONFIS). This F-CONFIS is a new type of a neural network whose links are with dependent and repeated weights between the input layer and hidden layer. For these special dependent repeated links of the F-CONFIS, some special properties are revealed. A new learning algorithm with these special properties is proposed in this paper for the F-CONFIS. The F-CONFIS is therefore applied for finding the capacity of the FNN. The lower bound and upper bound of the capacity of the FNN can be found from a new theorem proposed in this paper. Several examples are illustrated with satisfactory simulation results for the capacity of the F-CONFIS (or the FNN). These include "within capacity training of the FNN," "over capacity training of the FNN," "training by increasing the capacity of the FNN," and "impact of the capacity of the FNN in clustering Iris Data." It is noted that the finding of the capacity of the F-CONFIS, or FNN, has its emerging values in all engineering applications using fuzzy neural networks. This is to say that all engineering applications using FNN should not exceed the capacity of the FNN to avoid unexpected results. The clustering of Iris data using FNN illustrated in this paper is one of the most relevant engineering applications in this regards.