Filters








21,853 Hits in 3.8 sec

A hybrid connectionist-symbolic approach to regular grammatical inference based on neural learning and hierarchical clustering [chapter]

R. Alquezar, A. Sanfeliu
1994 Lecture Notes in Computer Science  
Several clustering algorithms have been suggested to extract a nite state automaton (FSA) from the activation patterns of a trained net.  ...  Recently, recurrent neural networks (RNNs) have been used to infer regular grammars from positive and negative examples.  ...  At the end of every neural run, a grammar is extracted in the form of a FSA.  ... 
doi:10.1007/3-540-58473-0_149 fatcat:wfwbotfxjngyfgpq7cdvdckt4a

Extracting and Learning an Unknown Grammar with Recurrent Neural Networks

C. Lee Giles, Clifford B. Miller, Dong Chen, Guo-Zheng Sun, Hsing-Hen Chen, Yee-Chun Lee
1991 Neural Information Processing Systems  
For many cases the extracted grammar outperforms the neural net from which it was extracted in correctly classifying unseen strings.  ...  After or during training. a dynamic clustering algorithm extracts the production rules that the neural network has learned..  ...  The University of Maryland authors gratefully acknowledge partial support from AFOSR and DARPA.  ... 
dblp:conf/nips/GilesMCSCL91 fatcat:nsxlujaqhbdyzll6xh3d47ss7a

Towards Interpretable ANNs: An Exact Transformation to Multi-Class Multivariate Decision Trees [article]

Duy T. Nguyen, Kathryn E. Kasmarik, Hussein A. Abbass
2021 arXiv   pre-print
In this paper, we introduce two novel multivariate decision tree (MDT) algorithms for rule extraction from ANNs: an Exact-Convertible Decision Tree (EC-DT) and an Extended C-Net algorithm.  ...  They both transform a neural network with Rectified Linear Unit activation functions into a representative tree, which can further be used to extract multivariate rules for reasoning.  ...  In EC-DT, the tree representation of the true process of the ANN, the rules extracted from the first hidden layer resemblance the same clustering method as implemented by Extended C-Net.  ... 
arXiv:2003.04675v4 fatcat:4ldfpgxwrfeivesgzjm5qyqipq

Learning and Extracting Finite State Automata with Second-Order Recurrent Neural Networks

C. L. Giles, C. B. Miller, D. Chen, H. H. Chen, G. Z. Sun, Y. C. Lee
1992 Neural Computation  
For a well-trained neural net, the extracted automata constitute an equivalence class of state machines that are reducible to the minimal machine of the inferred grammar.  ...  We show that a recurrent, second-order neural network using a realtime, forward training algorithm readily learns to infer small regular grammars from positive and negative string training samples.  ...  From the extracted FSA, minimal or not, the production rules of the learned grammar are evident.There are some interesting aspects to the extracted FSA.  ... 
doi:10.1162/neco.1992.4.3.393 fatcat:45z4b2matrgf7odd3y7mc6k5qi

Data mining of patients on weaning trials from mechanical ventilation using cluster analysis and neural networks

C. Arizmendi, E. Romero, R. Alquezar, P. Caminal, I. Diaz, S. Benito, B.F. Giraldo
2009 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society  
Applying a cluster analysis two groups with the majority dataset were obtained. Neural networks were applied to discriminate between patients from groups S, F and R.  ...  Each patient was characterized using 8 time series and 6 statistics extracted from respiratory and cardiac signals.  ...  Data mining is the process of extracting hidden patterns from data.  ... 
doi:10.1109/iembs.2009.5332742 pmid:19963824 fatcat:aan5cdmwmvedpgnksvljouwusq

Logic Explained Networks [article]

Gabriele Ciravegna, Pietro Barbiero, Francesco Giannini, Marco Gori, Pietro Lió, Marco Maggini, Stefano Melacci
2021 arXiv   pre-print
Experimental results on several datasets and tasks show that LENs may yield better classifications than established white-box models, such as decision trees and Bayesian rule lists, while providing more  ...  The large and still increasing popularity of deep learning clashes with a major limit of neural network architectures, that consists in their lack of capability in providing human-understandable motivations  ...  Here the authors focus on local rules extracted through input perturbation.  ... 
arXiv:2108.05149v1 fatcat:iovnkfoofzg35c7qv4y3wtp2na

A Cascaded Zoom-In Network for Patterned Fabric Defect Detection [article]

Zhiwei Zhang
2021 arXiv   pre-print
In the CZI-Net, the Aggregated HOG (A-HOG) and SIFT features are used to instead of simple convolution filters for feature extraction.  ...  Moreover, in order to extract more distinctive features, the feature representation layer and full connection layer are included in the CZI-Net.  ...  In the training state, all the local descriptors extracted from normal fabrics are clustered by K-Means algorithm to obtain the dictionary.  ... 
arXiv:2108.06760v2 fatcat:f6a4omvyprh3tht3axznc5khhe

Chromosome Extraction Based on U-Net and YOLOv3

Hua Bai, Tianhang Zhang, Changhao Lu, Wei Chen, Fangyun Xu, Zhibo Han
2020 IEEE Access  
Further, U-Net was used again to extract the single chromosomes precisely.  ...  The accuracy of extracting chromosomes from the raw G-band chromosome images reaches 99.3%. This method is of great significance for the development of automatic karyotype analysis technology.  ...  [14] extracted the chromosomes in the form of monomers and clusters using edge detection and the local threshold method. Shen X et al.  ... 
doi:10.1109/access.2020.3026483 fatcat:o3lsdsjhjngfxkvk5am6wcsoca

Healthcare knowledge of relationship between time series electrocardiogram and cigarette smoking using clinical records

Kuo-Kun Tseng, Jiaqian Li, Yih-Jing Tang, Ching Wen Yang, Fang-Ying Lin
2020 BMC Medical Informatics and Decision Making  
In this diagnostic system, several neural network models have been obtained from the different training subsets by clustering analysis.  ...  In this research, a combination of two techniques of pattern recognition; feature extraction and clustering neural networks, is specifically investigated during the diagnostic classification of cigarette  ...  Here, two simple and useful rules to combine all classifications from different neural network nets are introduced. One is the majority rule, and another is finding the best learning subset.  ... 
doi:10.1186/s12911-020-1107-2 pmid:32646409 fatcat:dsabqnprzfbqtjdw2pndfry4ui

Deep learning approaches in flow visualization

Can Liu, Ruike Jiang, Datong Wei, Changhe Yang, Yanda Li, Fang Wang, Xiaoru Yuan
2022 Advances in Aerodynamics  
Acknowledgements We would like to thank the funding support from NNW2018-ZT6B12 and NSFC No. 61872013, and the reviewers' feedback.  ...  Previous deep neural networks did not achieve comparable speed as local methods, as they suffer from several drawbacks.  ...  [24] propose Vortex-Net that uses the convolutional neural network to classify whether the central point of a local patch is inside a vortex structure.  ... 
doi:10.1186/s42774-022-00113-1 fatcat:ol5n6brdxbeu7nklyjdvgl7qo4

Using neural networks for data mining

Mark W. Craven, Jude W. Shavlik
1997 Future generations computer systems  
The rst type of approach, often called rule extraction, i n volves extracting symbolic models from trained neural networks.  ...  Speci cally, w e discuss two classes of approaches for data mining with neural networks.  ...  The local approach to rule extraction. A m ulti-layer neural network is decomposed into a set of single layer networks.  ... 
doi:10.1016/s0167-739x(97)00022-8 fatcat:hhcf2iai7nbm7hz6yyxpxjcsv4

Species Classification of Aquatic Plants using GRNN and ANFIS

S. Abirami, V. Ramalingam, S. Palanivel
2012 International Journal of Computer Applications  
Color, texture and shape features are extracted from 400 images of flowers.  ...  K-means clustering is used to extract the color features. Texture segmentation is done using texture filters. Edge detectors are used to trace the boundary of the image and hence the shape features.  ...  ANFIS approach learns the rules and membership functions. Rules: The if-then rules have to be determined somehow. This is usually done by "knowledge acquisition" from an expert.  ... 
doi:10.5120/7180-9863 fatcat:xf65idgxezfndh3ifgrnjg3i3e

Quality of Services Provisioning in Wireless Sensor Networks using Artificial Neural Network: A Survey

Mohit Mittal, Krishan Kumar
2015 International Journal of Computer Applications  
., while supervised learning includes perceptron model, delta learning rule, error back-propagation etc.  ...  In this paper we have surveyed artificial neural network for different QOS parameters of WSN. Artificial neural network (ANN) is very prominent emerging area for WSN applications.  ...  clusters while cluster head of each cluster play the role of a local base station [37] .  ... 
doi:10.5120/20553-2931 fatcat:fuaej2kk5bavphxena7w5xvuzy

Unsupervised Control Paradigm for Performance Evaluation

Sathya Ramadass, Annamma Abraham
2012 International Journal of Computer Applications  
There are different control paradigms available in the literature including Artificial Neural Networks, Fuzzy Logic Systems, Genetic Algorithms, Hybrid Models and others.  ...  Rule Extraction From Local Cluster Neural Nets. Neurocomputing. [12] Steiner, M.T.A., Neto, P.J.S., Soma, N.Y., Shimizu, T., and Nievola, J.C. 2006.  ...  In cluster approach, rule extraction is performed from a SOM by discovering clusters instead of boundaries [12] .  ... 
doi:10.5120/6380-8850 fatcat:kygxszpot5eyrjk6nabxtkke6m

Page 610 of American Society of Civil Engineers. Collected Journals Vol. 117, Issue CO4 [page]

1991 American Society of Civil Engineers. Collected Journals  
They could extract classification (clustering) characteristics from a large number of input examples, as in the case of unsupervised learning.  ...  This is due to the parallel structure of neural net- works. 4.  ... 
« Previous Showing results 1 — 15 out of 21,853 results