Search
Abstract
The paper analyzes some properties of denoising autoencoders using the problem of misspellings correction as an exemplary task. We evaluate the capacity of the network in its classical feed-forward form. We also propose a modification to the output layer of the net, which we called multi-softmax. Experiments show that the model trained with this output layer outperforms traditional network both in learning time and accuracy. We test the influence of the noise introduced to training data on the learning speed and generalization quality. The proposed approach of evaluating various properties of autoencoders using misspellings correction task serves as an open framework for further experiments, e.g. incorporating other neural network topologies into an autoencoder setting.
Citations
-
2
CrossRef
-
0
Web of Science
-
3
Scopus
Authors (2)
Cite as
Full text
- Publication version
- Accepted or Published Version
- License
- Copyright (Springer International Publishing AG 2017)
Keywords
Details
- Category:
- Conference activity
- Type:
- materiały konferencyjne indeksowane w Web of Science
- Title of issue:
- 9th International Conference on Computational Collective Intelligence (ICCCI) strony 438 - 447
- Language:
- English
- Publication year:
- 2017
- Bibliographic description:
- Draszawka K., Szymański J..: Analysis of Denoising Autoencoder Properties Through Misspelling Correction Task, W: 9th International Conference on Computational Collective Intelligence (ICCCI), 2017, ,.
- DOI:
- Digital Object Identifier (open in new tab) 10.1007/978-3-319-67077-5_42
- Verified by:
- Gdańsk University of Technology
seen 194 times
Recommended for you
Neural network agents trained by declarative programming tutors
- J. Dobrosolski,
- J. Szymański,
- H. Mora
- + 1 authors
Condition-Based Monitoring of DC Motors Performed with Autoencoders
- K. Włódarczak,
- Ł. Grzymkowski,
- T. Stefański