scholarly journals Improved Feature Learning: A Maximum-Average-Out Deep Neural Network for the Game Go

2020 ◽  
Vol 2020 ◽  
pp. 1-6
Author(s):  
Xiali Li ◽  
Zhengyu Lv ◽  
Bo Liu ◽  
Licheng Wu ◽  
Zheng Wang
Keyword(s):  
Neural Network ◽  
Low Cost ◽  
Computer Game ◽  
Feature Learning ◽  
Superior Performance ◽  
Computing Power ◽  
Unit Structure ◽  
The Neural Network ◽  
Neural Network Structure ◽  
Hidden Layer

Computer game-playing programs based on deep reinforcement learning have surpassed the performance of even the best human players. However, the huge analysis space of such neural networks and their numerous parameters require extensive computing power. Hence, in this study, we aimed to increase the network learning efficiency by modifying the neural network structure, which should reduce the number of learning iterations and the required computing power. A convolutional neural network with a maximum-average-out (MAO) unit structure based on piecewise function thinking is proposed, through which features can be effectively learned and the expression ability of hidden layer features can be enhanced. To verify the performance of the MAO structure, we compared it with the ResNet18 network by applying them both to the framework of AlphaGo Zero, which was developed for playing the game Go. The two network structures were trained from scratch using a low-cost server environment. MAO unit won eight out of ten games against the ResNet18 network. The superior performance of the MAO unit compared with the ResNet18 network is significant for the further development of game algorithms that require less computing power than those currently in use.

Building Electronic Commerce Recommendation System Based on Ontology Learning and BP Neural Network

2013 ◽  
Vol 718-720 ◽  
pp. 1961-1966
Author(s):  
Hong Sheng Xu ◽  
Qing Tan
Keyword(s):  
Neural Network ◽  
Electronic Commerce ◽  
Bp Neural Network ◽  
Optimization Problem ◽  
Recommendation System ◽  
High Efficiency ◽  
Ontology Learning ◽  
The Neural Network ◽  
Network Process ◽  
Hidden Layer

Electronic commerce recommendation system can effectively retain user, prevent users from erosion, and improve e-commerce system sales. BP neural network using iterative operation, solving the weights of the neural network and close values to corresponding network process of learning and memory, to join the hidden layer nodes of the optimization problem of adjustable parameters increase. Ontology learning is the use of machine learning and statistical techniques, with automatic or semi-automatic way, from the existing data resources and obtaining desired body. The paper presents building electronic commerce recommendation system based on ontology learning and BP neural network. Experimental results show that the proposed algorithm has high efficiency.

scholarly journals Advanced thyroid carcinomas: neural network analysis of ultrasonographic characteristics

2021 ◽  
Vol 14 (1) ◽  
Author(s):  
Michael Cordes ◽  
Theresa Ida Götz ◽  
Elmar Wolfgang Lang ◽  
Stephan Coerper ◽  
Torsten Kuwert ◽  
...  
Keyword(s):  
Neural Network ◽  
Imaging Modality ◽  
Follicular Carcinoma ◽  
Predictive Values ◽  
Thyroid Carcinomas ◽  
The Neural Network ◽  
Group 2 ◽  
Line Imaging ◽  
Hidden Layer ◽  
Group 1

Abstract Background Ultrasound is the first-line imaging modality for detection and classification of thyroid nodules. Certain characteristics observable by ultrasound have recently been identified that may indicate malignancy. This retrospective cohort study was conducted to test the hypothesis that advanced thyroid carcinomas show distinctive clinical and sonographic characteristics. Using a neural network model as proof of concept, nine clinical/sonographic features served as input. Methods All 96 study enrollees had histologically confirmed thyroid carcinomas, categorized (n = 32, each) as follows: group 1, advanced carcinoma (ADV) marked by local invasion or distant metastasis; group 2, non-advanced papillary carcinoma (PTC); or group 3, non-advanced follicular carcinoma (FTC). Preoperative ultrasound profiles were obtained via standardized protocols. The neural network had nine input neurons and one hidden layer. Results Mean age and the number of male patients in group 1 were significantly higher compared with groups 2 (p = 0.005) or 3 (p <  0.001). On ultrasound, tumors of larger volume and irregular shape were observed significantly more often in group 1 compared with groups 2 (p <  0.001) or 3 (p ≤ 0.01). Network accuracy in discriminating advanced vs. non-advanced tumors was 84.4% (95% confidence interval [CI]: 75.5–91), with positive and negative predictive values of 87.1% (95% CI: 70.2–96.4) and 92.3% (95% CI: 83.0–97.5), respectively. Conclusions Our study has shown some evidence that advanced thyroid tumors demonstrate distinctive clinical and sonographic characteristics. Further prospective investigations with larger numbers of patients and multicenter design should be carried out to show whether a neural network incorporating these features may be an asset, helping to classify malignancies of the thyroid gland.

scholarly journals ON SOME BOUNDS OF THE MINIMUM EDGE DOMINATING ENERGY OF A GRAPH

2021 ◽  
Vol 12 (4) ◽  
pp. 1285-1296
Author(s):  
Dr. Gauri Ghule , Et. al.
Keyword(s):  
Neural Network ◽  
Network Architecture ◽  
Fuzzy Controller ◽  
Back Propagation ◽  
Back Propagation Neural Network ◽  
Superior Performance ◽  
Fuzzy Classifier ◽  
Manual Intervention ◽  
The Neural Network ◽  
Hidden Neurons

Number of hidden neurons is necessary constant for tuning the neural network to achieve superior performance. These parameters are set manually through experimentation. The performance of the network is evaluated repeatedly to choose the best input parameters.Random selection of hidden neurons may cause underfitting or overfitting of the network. We propose a novel fuzzy controller for finding the optimal value of hidden neurons automatically. The hybrid classifier helps to design competent neural network architecture, eliminating manual intervention for setting the input parameters. The effectiveness of tuning the number of hidden neurons automatically on the convergence of a back-propagation neural network, is verified on speech data. The experimental outcomes demonstrate that the proposed Neuro-Fuzzy classifier can be viably utilized for speech recognition with maximum classification accuracy.

Neuron Optimization Based PID Approach for Cutting Force Control

2006 ◽  
Vol 315-316 ◽  
pp. 85-89
Author(s):  
S. Jiang ◽  
Yan Shen Xu ◽  
J. Wu
Keyword(s):  
Neural Network ◽  
Controller Design ◽  
Open Loop ◽  
Network Control ◽  
Neural Network Control ◽  
System A ◽  
The Neural Network ◽  
Nc Milling ◽  
Hidden Layer ◽  
Control Principle

To improve the cutting efficiency, one of key approaches is to control with constant force in the full depth working condition. And the controller design is vital to realize the real-time feasibility and robustness of the system. A neuron optimization based PID approach is proposed in this paper and adopted in the NC cutting process. This approach optimizes the parameters of PID controller real-timely with the neural network control principle. It not only overcomes the mismatch of the open-loop system model which occurred in constant PID control, but also solves the contradiction between the calculation speed and precision in the neural network which caused by the node choosing of the hidden layer. At last, the simulation has been carried out on a NC milling machine to prove the validity and effectiveness of the proposed approach.

scholarly journals Mind-Wave Controlled Robot: An Arduino Robot Simulating the Wheelchair for Paralyzed Patients

2018 ◽  
Vol 1 (1) ◽  
pp. 6
Author(s):  
Chi Hang Cheng ◽  
Shuai Li ◽  
Seifedine Kadry
Keyword(s):  
Neural Network ◽  
Low Cost ◽  
Eeg Signal ◽  
The Neural Network ◽  
Lead Electrode ◽  
Short Period ◽  
Specific Position ◽  
And Control

This project attempts to implement an Arduino robot to simulate a brainwave-controlled wheelchair for paralyzed patients with an improved controlling method. The robot should be able to move freely in anywhere under the control of the user and it is not required to predefine any map or path. An accurate and natural controlling method is provided, and the user can stop the robot any time immediately to avoid risks or danger. This project is using a low-cost and a brainwave-reading headset which has only a single lead electrode (Neurosky mind wave headset) to collect the EEG signal. BCI will be developed by sending the EEG signal to the Arduino Mega and control the movement of the robot. This project used the eye blinking as the robot controlling method as the eye blinking will cause a significant pulse in the EEG signal. By using the neural network to classify the blinking signal and the noise, the user can send the command to control the robot by blinking twice in a short period of time. The robot will be evaluated by driving in different places to test whether it can follow the expected path, avoid the obstacles, and stop in a specific position.

scholarly journals RGB-D Object Recognition Using Multi-Modal Deep Neural Network and DS Evidence Theory

Sensors ◽  
2019 ◽  
Vol 19 (3) ◽  
pp. 529 ◽  
Author(s):  
Hui Zeng ◽  
Bin Yang ◽  
Xiuqing Wang ◽  
Jiwei Liu ◽  
Dongmei Fu
Keyword(s):  
Neural Network ◽  
Object Recognition ◽  
Deep Neural Network ◽  
Low Cost ◽  
Evidence Theory ◽  
Feature Learning ◽  
Decision Fusion ◽  
Support Vector ◽  
Depth Sensors ◽  
Depth Images

With the development of low-cost RGB-D (Red Green Blue-Depth) sensors, RGB-D object recognition has attracted more and more researchers’ attention in recent years. The deep learning technique has become popular in the field of image analysis and has achieved competitive results. To make full use of the effective identification information in the RGB and depth images, we propose a multi-modal deep neural network and a DS (Dempster Shafer) evidence theory based RGB-D object recognition method. First, the RGB and depth images are preprocessed and two convolutional neural networks are trained, respectively. Next, we perform multi-modal feature learning using the proposed quadruplet samples based objective function to fine-tune the network parameters. Then, two probability classification results are obtained using two sigmoid SVMs (Support Vector Machines) with the learned RGB and depth features. Finally, the DS evidence theory based decision fusion method is used for integrating the two classification results. Compared with other RGB-D object recognition methods, our proposed method adopts two fusion strategies: Multi-modal feature learning and DS decision fusion. Both the discriminative information of each modality and the correlation information between the two modalities are exploited. Extensive experimental results have validated the effectiveness of the proposed method.

Sunlight Greenhouse Temperature Prediction Model Based on Bayesian Regularization BP Neural Network

2015 ◽  
Vol 740 ◽  
pp. 871-874
Author(s):  
Hui Zhao ◽  
Li Rong Shi ◽  
Hong Jun Wang
Keyword(s):  
Neural Network ◽  
Network Structure ◽  
Bp Neural Network ◽  
Principal Component ◽  
Coupling Factor ◽  
Bayesian Regularization ◽  
The Neural Network ◽  
Neural Network Structure ◽  
Input Variables ◽  
Bp Neural Network Model

Directing against the problems of too large size of the neural network structure due to the existence of a complex relationship between the input coupling factor and too many input factors in establishing model for predicting temperature of sunlight greenhouse. This article chose the environmental factors that affect the sunlight greenhouse temperature as data sample. Through the principal component analysis of data samples, three main factors were extracted. These selected principal component values were taken as the input variables of BP neural network model. Use the Bayesian regularization algorithm to improve the BP neural network. The empirical results show that this method is utilized modify BP neural network, which can simplify network structure and smooth fitting curve, has good generalization capability.

scholarly journals Towards Better Interpretability in Deep Q-Networks

2019 ◽  
Vol 33 ◽  
pp. 4561-4569 ◽  
Author(s):  
Raghuram Mandyam Annasamy ◽  
Katia Sycara
Keyword(s):  
Neural Network ◽  
Network Architecture ◽  
Empirical Studies ◽  
Superior Performance ◽  
Training Algorithms ◽  
Neural Network Architecture ◽  
Q Learning ◽  
The Neural Network ◽  
Learning Techniques ◽  
Out Of Sample

Deep reinforcement learning techniques have demonstrated superior performance in a wide variety of environments. As improvements in training algorithms continue at a brisk pace, theoretical or empirical studies on understanding what these networks seem to learn, are far behind. In this paper we propose an interpretable neural network architecture for Q-learning which provides a global explanation of the model’s behavior using key-value memories, attention and reconstructible embeddings. With a directed exploration strategy, our model can reach training rewards comparable to the state-of-the-art deep Q-learning models. However, results suggest that the features extracted by the neural network are extremely shallow and subsequent testing using out-of-sample examples shows that the agent can easily overfit to trajectories seen during training.

scholarly journals Prediction of Slag Characteristics Based on Artificial Neural Network for Molten Gasification of Hazardous Wastes

Energies ◽  
2020 ◽  
Vol 13 (19) ◽  
pp. 5115
Author(s):  
Xiongchao Lin ◽  
Wenshuai Xi ◽  
Jinze Dai ◽  
Caihong Wang ◽  
Yonggang Wang
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Layer Structure ◽  
Hazardous Materials ◽  
Hazardous Wastes ◽  
Slag Viscosity ◽  
The Neural Network ◽  
Independent Test ◽  
Artificial Neural ◽  
Hidden Layer

Molten gasification is considered as a promising technology for the processing and safe disposal of hazardous wastes. During this process, the organic components are completely converted while the hazardous materials are safely embedded in slag via the fusion-solidification-vitrification transformation. Ideally, the slag should be glassy with low viscosity to ensure the effective immobilization and steady discharge of hazardous materials. However, it is very difficult to predict the characteristics of slag using existing empirical equations or conventional mathematical methods, due to the complex non-linear relationship among the phase transformation, vitrification transition and chemical composition of slag. Equipped with a strong nonlinear mapping ability, an artificial neural network may be able to predict the properties of slags if a large amount of data is available for training. In this work, over 10,000 experimental data points were used to train and develop a slag classification model (glassy vs. non-glassy) based on a neural network. The optimal structure of the neural network was figured out and validated. The results suggest that the classification accuracy for the independent test samples reached 93.3%. Using 1 and 0 as model inputs to represent mildly reducing and inert atmospheres, a double hidden layer structure in the neural network enabled the accurate classification of slags under various atmospheres. Furthermore, the neural network for the prediction of glassy slag viscosity was optimized; it featured a double hidden layer structure. Under a mildly reducing atmosphere, the absolute error from the independent test data was generally within 4 Pa·s. By adding a gas atmosphere into the input of the neural network using a simple normalization method, a multi-atmosphere slag viscosity prediction model was developed. Said model is much more accurate than its counterpart that does not consider the effect of the atmosphere. In summary, the artificial neural network proved to be an effective approach to predicting the slag properties under different atmospheres. The data-driven models developed in this work are expected to facilitate the commercial deployment of molten gasification technology.

STDP Provides the Substrate for Igniting Synfire Chains by Spatiotemporal Input Patterns

2008 ◽  
Vol 20 (2) ◽  
pp. 415-435 ◽  
Author(s):  
Ryosuke Hosaka ◽  
Osamu Araki ◽  
Tohru Ikeguchi
Keyword(s):  
Neural Network ◽  
Experimental Studies ◽  
Spike Timing ◽  
Inhibitory Neurons ◽  
Spatiotemporal Pattern ◽  
Spiking Neural Network ◽  
Synfire Chains ◽  
The Neural Network ◽  
Stdp Rule ◽  
Neural Network Structure

Spike-timing-dependent synaptic plasticity (STDP), which depends on the temporal difference between pre- and postsynaptic action potentials, is observed in the cortices and hippocampus. Although several theoretical and experimental studies have revealed its fundamental aspects, its functional role remains unclear. To examine how an input spatiotemporal spike pattern is altered by STDP, we observed the output spike patterns of a spiking neural network model with an asymmetrical STDP rule when the input spatiotemporal pattern is repeatedly applied. The spiking neural network comprises excitatory and inhibitory neurons that exhibit local interactions. Numerical experiments show that the spiking neural network generates a single global synchrony whose relative timing depends on the input spatiotemporal pattern and the neural network structure. This result implies that the spiking neural network learns the transformation from spatiotemporal to temporal information. In the literature, the origin of the synfire chain has not been sufficiently focused on. Our results indicate that spiking neural networks with STDP can ignite synfire chains in the cortices.

Sign in / Sign up

Export Citation Format

Share Document