Update state of the art

[JJ]

After #18 update that section.

[pacastillo]

Maybe we could include reference to the 2-fold or 10-fold cross-validation methods that are used in the ANN field.

Also to the way that some authors propose take advantage of uncertainty (see below).

---

There are some approaches to deal with uncertainty in ANN, such as 2-fold or 10-fold cross-validation [Haykin1999, ALPTEKIN2013].

Also, there are some approaches that take advantage of uncertainty. In example in [Ligomenides1993], the authors suggest that uncertainty may be managed naturally, and even used profitably, in cooperative, self-organizing, dynamical physical systems, and in neural networks. Then, in [Blundell2015], the authors demonstrate how the learnt uncertainty in the weights of a ANN can be used to improve generalisation in non-linear regression problems, and how this weight uncertainty can be used to drive the exploration-exploitation trade-off in reinforcement learning.

---

Haykin, S. Neural Networks: A Comprehensive Foundation (second edn). Prentice-Hall Inc., Englewood Cliffs, New Jersey, USA. 1999

AHMET ALPTEKIN and OLCAY KURSUN MISS ONE OUT: A CROSS-VALIDATION METHOD UTILIZING INDUCED TEACHER NOISE. Int. J. Patt. Recogn. Artif. Intell. 27, 1351003 (2013) [11 pages] DOI: 10.1142/S0218001413510038

P.A. Ligomenides Uncertainty in neural networks Proceedings of the Second International Symposium on Uncertainty Modeling and Analysis, Page(s) 83 - 89. 1993.

Charles Blundell, Julien Cornebise, Koray Kavukcuoglu, Daan Wierstra Weight Uncertainty in Neural Networks Proceedings of the 32nd International Conference on Machine Learning, Lille, France, 2015. JMLR: W&CP volume 37. arXiv:1505.05424v2 [stat.ML] 21 May 2015

[pacastillo]

In the filed of ANN, the one-to many mapping from genotype to the actual networks (phenotypes) may induce noisy fitness evaluation [Jagtap2014]. In this paper Jagtap et al. propose an improved QNN method to provide a set of appropriate weights when evolving the network structure and to alleviate the noisy fitness evaluation problem.

Avinash Jagtap, Rashmi Deshpande Efficient Method for Optimizing Artificial Neural Network Using Quantum-Based Algorithm International Journal of Advanced Research in Computer Science and Software Engineering Volume 4, Issue 6, pp. 692-700. June 2014 ISSN: 2277 128X

[pacastillo]

Also, we should reference the following one:

In [Flores2011] a new Rank Based Selection operator and a new variation of a Rank Based Mutation are proposed to evolve the neural network topology used as a controller for a robot. The operators were suitable for this kind of problems where the fitness landscape is noisy.

---

@inproceedings{Flores2011, title={Rank Based Evolution of Real Parameters on Noisy Fitness Functions: Evolving a Robot Neurocontroller}, author={D. Flores}, booktitle={10th Mexican International Conference on Artificial Intelligence (MICAI)}, pages={72--76}, year={2011}, organization={IEEE}, doi={10.1109/MICAI.2011.40} }

[JJ]

Please add them to the .bib file and I'll take a look at then and update the state of the art.

2016-01-17 19:19 GMT+01:00 Pedro A. Castillo Valdivieso < notifications@github.com>:

Also, we should reference the following one:

In [Flores2011] a new Rank Based Selection operator and a new variation of a Rank Based Mutation are proposed to evolve the neural network topology used as a controller for a robot. The operators were suitable for this kind

of problems where the fitness landscape is noisy.

@inproceedings{Flores2011, title={Rank Based Evolution of Real Parameters on Noisy Fitness Functions: Evolving a Robot Neurocontroller}, author={D. Flores}, booktitle={10th Mexican International Conference on Artificial Intelligence (MICAI)}, pages={72--76}, year={2011}, organization={IEEE}, doi={10.1109/MICAI.2011.40} }

— Reply to this email directly or view it on GitHub https://github.com/geneura-papers/2015-ECTA/issues/19#issuecomment-172360287 .

JJ