International Conference on Hybrid Artificial Intelligence Systems

Conference Information
Name: International Conference on Hybrid Artificial Intelligence Systems
Acronym: HAIS
Location: Bilbao, Spain
Date: 22 September 2021 - 24 September 2021

Articles from this conference

Alessandro Baldo, , Edoardo Fadda, Pablo G. Bringas
Published: 15 September 2021
Disinformation in Open Online Media pp 550-563; https://doi.org/10.1007/978-3-030-86271-8_46

Abstract:
In this study, we propose a deeper analysis on the algorithmic treatment of financial time series, with a focus on Forex markets’ applications. The relevant aspects of the paper refers to a more beneficial data arrangement, proposed into a two-dimensional objects and to the application of a Temporal Convolutional Neural Network model, representing a more than valid alternative to Recurrent Neural Networks. The results are supported by expanding the comparison to other more consolidated deep learning models, as well as with some of the most performing Machine Learning methods. Finally, a financial framework is proposed to test the real effectiveness of the algorithms.
, Alberto Gallucci, Jose Ramón Villar, Kaori Yoshida, Mario Koeppen
Published: 15 September 2021
Disinformation in Open Online Media pp 659-670; https://doi.org/10.1007/978-3-030-86271-8_55

Abstract:
A Speech Emotion Recognition (SER) system can be defined as a collection of methodologies that process and classify speech signals to detect emotions embedded in them [2]. Among the most critical issues to consider in an SER system are: i) definition of the kind of emotions to classify, ii) look for suitable datasets, iii) selection of the proper input features and iv) optimisation of the convenient features. This work will consider four of the well-known dataset in the literature: EmoDB, TESS, SAVEE and RAVDSS. Thus, this study focuses on designing a low-power SER algorithm based on combining one prosodic feature with six spectral features to capture the rhythm and frequency. The proposal compares eleven low-power Classical classification Machine Learning techniques (CML), where the main novelty is optimising the two main parameters of the MFCC spectral feature through the meta-heuristic technique SA: the n_mfcc and the hop_length. The resulting algorithm could be deployed on low-cost embedded systems with limited computational power like a smart speaker. In addition, the proposed SER algorithm will be validated for four well-known SER datasets. The obtained models for the eleven CML techniques with the optimised MFCC features outperforms clearly (more than a 10%) the baseline models obtained with the not-optimised MFCC for the studied datasets.
Santiago Porras, , Bruno Baruque, José Luis Calvo-Rolle
Published: 15 September 2021
Disinformation in Open Online Media pp 500-510; https://doi.org/10.1007/978-3-030-86271-8_42

Abstract:
The environmental impact caused by greenhouse gasses emissions derived from fossil fuels, gives rise to the promotion of green policies. In this context, geothermal energy systems has experienced a significant increase in its use. The efficiency of this technology is closely linked with factors such as ground temperature, weather and season. This work develops the analysis of the behaviour of a geothermal system placed in a bioclimatic house during one year, by means of projectionists and clustering methods.
, José-Luis Casteleiro-Roca, Dragan Simić, José Luis Calvo-Rolle
Published: 15 September 2021
Disinformation in Open Online Media pp 367-378; https://doi.org/10.1007/978-3-030-86271-8_31

Abstract:
In this research, a study about the implementation of a hybrid intelligent model for classification applied to power electronics is presented. First of all, an analysis of the chosen power converter, half-bridge buck converter, has been done, differentiating between two operating modes: Hard-Switching and Soft-Switching. A hybrid model combining a clustering method with classification intelligent techniques is implemented. The obtained model differentiate with high accuracy between the two modes, obtaining very good results in the classification.
Published: 15 September 2021
Disinformation in Open Online Media pp 196-206; https://doi.org/10.1007/978-3-030-86271-8_17

Abstract:
The increasing implementation of software and data analysis tools based on Artificial Intelligence in the field of law, raises technical and ethical questions regarding the legal nature of the actions that take place in an area of ​​eJustice. In the first place, the ethical problems derived from Jurimetrics appear: the increasing capacity to detect hidden patterns between the data of sentences and procedural forms, allow establishing statistical guidelines of reliability in terms of the application of a certain line of defense or another. This endangers the duty of every legal agent (not only of the judges) to ensure justice, turning the orientations and the choices of argumentative strategies into a commercial practice that distorts the ethical nature of the legal profession. On the other hand, the irruption of new communication technologies and data analysis can modify the conditions of establishment and development of both the procedures and the judicial process itself (both civil and criminal). Finally, AI confronts us with a series of ethical problems derived from the predictive function applied to research. This work will try to briefly show these issues and conclude with a series of ethical indications or recommendations to avoid dangers and turn digitization into a tool at the service of the legal operator and at the service of defending the rights and dignity of people.
, Iván Sevillano-García, María Jesús Lucena-González, José Luis Martín-Rodríguez, ,
Published: 15 September 2021
Disinformation in Open Online Media pp 305-315; https://doi.org/10.1007/978-3-030-86271-8_26

Abstract:
Datasets from real-world applications usually deal with many variables and present difficulties when modeling them with traditional classifiers. There is a variety of feature selection and extraction tools that may help with the dimensionality problem, but most of them do not focus on the complexity of the classes. In this paper, a new autoencoder-based model for addressing class complexity in data is introduced, aiming to extract features that present classes in a more separable fashion, thus simplifying the classification task. This is possible thanks to a combination of the standard reconstruction error with a least-squares support vector machine loss function. This model is then applied to a practical use case: classification of chest X-rays according to the presence of COVID-19, showing that learning features that increase linear class separability can boost classification performance. For this purpose, a specific convolutional autoencoder architecture has been designed and trained using the recently published COVIDGR dataset. The proposed model is evaluated by means of several traditional classifiers and metrics, in order to establish the improvements caused by the extracted features. The advantages of using a feature learner and traditional classifiers are also discussed.
Back to Top Top