| viXra.org > Artificial Intelligence > 2202 |
 |
 |
| viXra.org > Artificial Intelligence > 2202 |
|
|
[7] viXra:2202.0162 [pdf] submitted on 2022-02-25 19:21:37
Authors: Siddhant Kumar Jha
Comments: 6 Pages.
The objective of the study is to develop a definitive meta-analysis of the recent developments in hyper-graph theories’ application in the field and study of deep learning and more widely in Machine learning , the applications of this particular technique may range simple classification tuning to more advanced abstract GANs in the field of regenerative graphical systems and computer vision in general,In our experiments, we use a novel random walk procedure and show that our model achieves and, in most cases, surpasses state-of-the-art performance on benchmark data sets. Additionally we also try to display our classification performance as
compared to traditional Statistical Techniques , ML algorithms as well as Classical and new Deep learning algorithms.
Category: Artificial Intelligence
[6] viXra:2202.0116 [pdf] replaced on 2022-04-12 05:22:42
Authors: Jeongik Cho
Comments: 8 Pages.
Dynamic latent scale GAN proposed a learning-based GAN inversion method with maximum likelihood estimation. In this paper, we propose a method for self-supervised out-of-distribution detection using the encoder of dynamic latent scale GAN. When the dynamic latent scale GAN converged, since the entropy of the scaled latent random variable is optimal to represent in-distribution data, in-distribution data is densely mapped to latent codes with high likelihood. This enables the log-likelihood of the predicted latent code to be used for out-of-distribution detection. The proposed method does not require mutual information of in-distribution data and additional hyperparameters for prediction. The proposed method also showed better out-of-distribution detection performance than the previous state-of-art method.
Category: Artificial Intelligence
[5] viXra:2202.0106 [pdf] replaced on 2022-06-04 10:23:09
Authors: Ait-Taleb Nabil
Comments: 26 Pages.
In this paper, we will expose the BIC score expressed as a function of the Bayesian network's entropy.
We will then use this BIC score to learn a Bayesian network from an example of data frame.
Category: Artificial Intelligence
[4] viXra:2202.0082 [pdf] submitted on 2022-02-14 01:43:59
Authors: Mihai Oltean, Dumitru Dumitrescu
Comments: 10 Pages. International Conference on Computational Sciences, ICCS'04, Edited by M. Bubak, G. D. van Albada, P. Sloot, and J. Dongarra, Vol. II, pp. 670-673, 6-9 June, Krakow, Poland, Springer-Verlag, Berlin, 2004.
Multi Expression Programming (MEP) is an evolutionary technique that may be used for solving computationally difficult problems. MEP uses a linear solution representation. Each MEP
individual is a string encoding complex expressions (computer programs). An MEP individual may encode multiple solutions of the current problem. In this paper, MEP is used for evolving a Traveling Salesman Problem (TSP) heuristic for graphs satisfying triangle inequality. Evolved MEP
heuristic is compared with Nearest Neighbor Heuristic (NN) and Minimum Spanning Tree Heuristic (MST) on some difficult problems in TSPLIB. For most of the considered problems the evolved MEP heuristic outperforms NN and MST. The obtained algorithm was tested against some problems in TSPLIB. The results emphasize that evolved MEP
heuristic is a powerful tool for solving difficult TSP instances.
Category: Artificial Intelligence
[3] viXra:2202.0081 [pdf] submitted on 2022-02-14 01:46:16
Authors: Mihai Oltean, Crina Grosan
Comments: NASA/DoD Conference on Evolvable Hardware, 24-26 June, Seattle, Edited by R. Zebulum (et. al), pages 87-90, IEEE Press, NJ, 2004
Multi Expression Programming (MEP) is a Genetic
Programming (GP) variant that uses linear chromosomes for solution encoding. A unique MEP feature is its ability of encoding multiple solutions of a problem in a single chromosome. These solutions are handled in the same time
complexity as other techniques that encode a single solution in a chromosome. In this paper, MEP is used for evolving digital circuits. MEP is compared to Cartesian Genetic Programming
(CGP) – a technique widely used for evolving digital circuits – by using several well-known problems in the field of electronic circuit design. Numerical experiments show that MEP
outperforms CGP for the considered test problems.
Category: Artificial Intelligence
[2] viXra:2202.0080 [pdf] submitted on 2022-02-14 01:49:14
Authors: Mihai Oltean
Comments: 4 Pages. Proceedings of the 5th International Workshop on Frontiers in Evolutionary Algorithms, The 7th Joint Conference on Information Sciences, September 26-30, 2003, Research Triangle Park, North Carolina, Edited by Ken Chen (et. al), pp. 315-318, 2003.
In this paper, the Multi Expression Programming
(MEP) technique is used for solving even-parity
problems. Numerical experiments show that MEP
outperforms Genetic Programming (GP) with more than
one order of magnitude for the considered test cases.
Category: Artificial Intelligence
[1] viXra:2202.0079 [pdf] submitted on 2022-02-14 01:51:37
Authors: Mihai Oltean
Comments: 36 Pages. chapter 10, Evolvable Machines: Theory and Applications, Springer-Verlag, edited by Nadia Nedjah (et al.), pp. 229-255, 2004
Multi Expression Programming is a Genetic Programming variant that uses a linear representation of individuals. A unique feature of Multi Expression Programming is its ability of storing multiple solutions of a problem in a single chromosome. In this paper, we propose and use several techniques for improving the search performed by Multi Expression Programming. Some of the most important improvements are Automatically Defined Functions and Sub-Symbolic node representation. Several experiments with Multi Expression Programming are performed in this paper. Numerical results show that Multi Expression Programming performs very well for the
considered test problems.
Category: Artificial Intelligence
| Third-Party Links: |
|