Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 14
Encontrada(s) 1 página(s)

UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Corpus for cyberbullying prevention

  • Moreno Bilbao, M. Asunción|||0000-0002-1823-5970
  • Bonafonte Cávez, Antonio|||0000-0002-6240-9915
  • Jauk, Igor
  • Tarrés, Laia
  • Pereira, Victor
Cyberbullying is the use of digital media to harass a person or group of people, through personal attacks, disclosure of confidential or false information, among other means. That is to say, it is considered cyberbullying, or cyber-aggression to everything that is done through electronic communication devices with the intended purpose of harming or attacking a person or a group.In this paper we present a starting project to prevent cyberbullying between kids and teenagers. The idea is to create a prevention system. A system which is installed in the mobile of a kid and, if a harassment is detected, some advice is given to the child. In case of serious or repeated behavior the parents are alerted. The focus of this paper is to describe the characteristics of the database to be used to train the system, Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Corpus for cyberbullying prevention

  • Moreno Bilbao, M. Asunción
  • Bonafonte Cávez, Antonio
  • Jauk, Igor
  • Tarrés, Laia
  • Pereira, Victor
Cyberbullying is the use of digital media to harass a person or group of people, through personal attacks, disclosure of confidential or false information, among other means. That is to say, it is considered cyberbullying, or cyber-aggression to everything that is done through electronic communication devices with the intended purpose of harming or attacking a person or a group.In this paper we present a starting project to prevent cyberbullying between kids and teenagers. The idea is to create a prevention system. A system which is installed in the mobile of a kid and, if a harassment is detected, some advice is given to the child. In case of serious or repeated behavior the parents are alerted. The focus of this paper is to describe the characteristics of the database to be used to train the system, Peer Reviewed, Postprint (published version)
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Emotion Detection from Speech and Text

  • De Velasco Vázquez, Mikel
  • Justo Blanco, Raquel
  • Antón, Josu
  • Carrilero, Mikel
  • Torres Barañano, María Inés
The main goal of this work is to carry out automatic emo-tion detection from speech by using both acoustic and textualinformation. For doing that a set of audios were extracted froma TV show were different guests discuss about topics of currentinterest. The selected audios were transcribed and annotatedin terms of emotional status using a crowdsourcing platform.A 3-dimensional model was used to define an specific emo-tional status in order to pick up the nuances in what the speakeris expressing instead of being restricted to a predefined set ofdiscrete categories. Different sets of acoustic parameters wereconsidered to obtain the input vectors for a neural network. Torepresent each sequence of words, a models based on word em-beddings was used. Different deep learning architectures weretested providing promising results, although having a corpus ofa limited size., This work has been partially founded by bythe Spanish Government (TIN2014-54288-C4-4-R and TIN2017-85854-C4-3-R), and bythe European Commission H2020 SC1-PM15program under RIA 7 grant 69872.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Capítulo o parte de un libro (BookPart)

Regularized Neural User Model for Goal-Oriented Spoken Dialogue Systems

  • Serras Sáenz, Manex
  • Torres Barañano, María Inés
  • Del Pozo, Arantza
User simulation is widely used to generate artificial dialogues in order to train statistical spoken dialogue systems and perform evaluations. This paper presents a neural network approach for user modeling that exploits an encoder-decoder bidirectional architecture with a regularization layer for each dialogue act. In order to minimize the impact of data sparsity, the dialogue act space is compressed according to the user goal. Experiments on the Dialogue State Tracking Challenge 2 (DSTC2) dataset provide significant results at dialogue act and slot level predictions, outperforming previous neural user modeling approaches in terms of F1 score., Spanish Minister of Science under grants TIN2014-54288-C4-4-R and TIN2017-85854-C4-3-R and by the EU H2020 EMPATHIC project grant number 769872.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Iruzurrezko portaeren detekzioa crowd motako etiketazioan

  • De Velasco Vázquez, Mikel
  • López Zorrilla, Asier
  • Justo Blanco, Raquel
This work aims at detecting low quality labels in crowdsourcing annotation tasks. We validate our proposal carrying out experiments in a difficult and subjective task: emotion recognition. We have developed several measures in order to detect fraudulent behaviour, including measures related to the labelling time, worker inter-agreement and the distribution of the answers. Not only do we show that each of the described measures is helpful but we also demonstrate that mixing them is the best way to go., Lan honek crowd motako etiketazioan agertu daitezkeen kalitate baxuko etiketak detektatzea du helburu. Proposatutako metodologia balioztatzeko, saiakuntzak ataza zail eta subjektibo batekin egin ditugu: emozioen de- tekzioarekin. Iruzurrezko langileak topatzeko zenbait neurri proposatu dira, etiketatze denboran, langileen arteko adostasunean eta langileen erantzunen banaketan oinarriturikoak. Neurri bakoitza baliagarria dela frogatu dugun arren, gure ondorio nagusia neurriak batzerakoan iruzurrezko langileak detektatzeko probabilitatea handitzen dela da., Egileok gure esker ona adierazi nahiko genioke Euskal Herriko Unibertsitateari, Espainako gobernuako TIN2017- 85854-C4-3-R zenbakidun diru laguntzari eta H2020 Europako Batzordeko SC1-PM15 programako RIA 7 deial- diko 769872 zenbakidun laguntzari, hurrenez hurren, ikerketa hau babesteagatik.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

A Differentiable Generative Adversarial Network for Open Domain Dialogue

  • López Zorrilla, Asier
  • De Velasco Vázquez, Mikel
  • Torres Barañano, María Inés
Presentado antes de su publicación por Springer en el IWSDS 2019: International Workshop on Spoken Dialogue Systems Technology, Siracusa, Italy, April 24-26, 2019, This work presents a novel methodology to train open domain neural dialogue systems within the framework of Generative Adversarial Networks with gradient-based optimization methods. We avoid the non-differentiability related to text-generating networks approximating the word vector corresponding to each generated token via a top-k softmax. We show that a weighted average of the word vectors of the most probable tokens computed from the probabilities resulting of the top-k softmax leads to a good approximation of the word vector of the generated token. Finally we demonstrate through a human evaluation process that training a neural dialogue system via adversarial learning with this method successfully discourages it from producing generic responses. Instead it tends to produce more informative and variate ones., This work has been partially funded by the Basque Government under grant PRE_2017_1_0357, by the University of the Basque Country UPV/EHU under grant PIF17/310, and by the H2020 RIA EMPATHIC (Grant N: 769872).
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Capítulo o parte de un libro (BookPart)

Tracking the Expression of Annoyance in Call Centers

  • Irastorza Manso, Jon
  • Torres Barañano, María Inés
Machine learning researchers have dealt with the identification of emo- tional cues from speech since it is research domain showing a large number of po- tential applications. Many acoustic parameters have been analyzed when searching for cues to identify emotional categories. Then classical classifiers and also out- standing computational approaches have been developed. Experiments have been carried out mainly over induced emotions, even if recently research is shifting to work over spontaneous emotions. In such a framework, it is worth mentioning that the expression of spontaneous emotions depends on cultural factors, on the particu- lar individual and also on the specific situation. In this work, we were interested in the emotional shifts during conversation. In particular we were aimed to track the annoyance shifts appearing in phone conversations to complaint services. To this end we analyzed a set of audio files showing different ways to express annoyance. The call center operators found disappointment, impotence or anger as expression of annoyance. However, our experiments showed that variations of parameters derived from intensity combined with some spectral information and suprasegmental fea- tures are very robust for each speaker and annoyance rate. The work also discussed the annotation problem arising when dealing with human labelling of subjective events. In this work we proposed an extended rating scale in order to include anno- tators disagreements. Our frame classification results validated the chosen annota- tion procedure. Experimental results also showed that shifts in customer annoyance rates could be potentially tracked during phone calls, Spanish Mineco under grant TIN2014- 54288-C4-4-R H2020 EU under Empathic RIA action number 769872.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Euskaraz hitz egiten ikasten duten makina autodidaktak

  • López Zorrilla, Asier
  • De Velasco Vázquez, Mikel
  • Justo Blanco, Raquel
Lan honetan sare neuronalen bidez euskaraz hitz egiten ikasten duen elkarrizketa sistema automatikoa aurkezten dugu. Horretarako, Turingen testaren ideia era konputazionalean inplementatzen duten sare neuronal sortzaile aurkariak erabili ditugu. Normalean erabiltzen diren ingelesezko corpusak baino bi magnitude ordena txikiagoa den euskarazko corpus batekin halako sareak entrenatzea badagoela frogatzen dugu. Amaitzeko, euskararen morfologia kontuan hartzen duen aurreprozesamendua erabiltzea komenigarria dela erakusten dugu. Dakigunaren arabera, sare neuronaletan oinarrituta dagoen euskarazko lehen elkarrizketa sistema aurkezten dugu., Lan honen egileok gure esker ona adierazi nahiko genioke Eusko Jaurlaritzari, Euskal Herriko Unibertsitateari eta baita Europar Batzordeari, PRE 2017 1 0357 eta PIF17/310 zenbakidun diru laguntzekin, eta H2020 SC1-PM15 programako RIA 7 deialdiko 769872 zenbakidun diru laguntzarekin, hurrenez hurren, ikerketa hau babesteagatik.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Can Spontaneous Emotions be Detected from Speech on TV Political Debates?

  • De Velasco Vázquez, Mikel
  • Justo Blanco, Raquel
  • López Zorrilla, Asier
  • Torres Barañano, María Inés
Accepted paper, Decoding emotional states from multimodal signals is an increasingly active domain, within the framework of affective computing, which aims to a better understanding of Human-Human Communication as well as to improve Human- Computer Interaction. But the automatic recognition of sponta- neous emotions from speech is a very complex task due to the lack of a certainty of the speaker states as well as to the difficulty to identify a variety of emotions in real scenarios. In this work we explore the extent to which emotional states can be decoded from speech signals extracted from TV political debates. The labelling procedure was supported by perception experiments where only a small set of emotions has been identified. In addition, some scaled judgements of valence, arousal and dominance were also provided. In this framework the paper shows meaningful comparisons between both, the dimensional and the categorical models of emotions, which is a new con- tribution when dealing with spontaneous emotions. To this end Support Vector Machines (SVM) as well as Feedforward Neural Networks (FNN) have been proposed to develop classifiers and predictors. The experimental evaluation over a Spanish corpus has shown the ability of both models to be identified in speech segments by the proposed artificial systems., This work has been partially funded by the Spanish Government under grant TIN2017-85854-C4-3-R (AEI/FEDER,UE) and conducted in the project EMPATHIC (Grant n769872) funded by the European Union’s H2020 research andinnovation program.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Adding dimensional features for emotion recognition on speech

  • Ben Letaifa Zouari, Leila
  • Torres Barañano, María Inés
  • Justo Blanco, Raquel
Developing accurate emotion recognition systems requires extracting suitable features of these emotions. In this paper, we propose an original approach of parameters extraction based on the strong, theoretical and empirical, correlation between the emotion categories and the dimensional emotions parameters. More precisely, acoustic features and dimensional emotion parameters are combined for better speech emotion characterisation. The procedure consists in developing arousal and valence models by regression on the training data and estimating, by classification, their values in the test data. Hence, when classifying an unknown sample into emotion categories, these estimations could be integrated into the feature vectors. It is noted that the results using this new set of parameters show a significant improvement of the speech emotion recognition performance., Spanish Minister of Science under grant TIN2017-85854-C4-3-R European Union’s Horizon 2020 research and innovation program under grant agreement No 769872
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

First Steps to Develop a Corpus of Interactions between Elderly and Virtual Agents in Spanish with Emotion Labels

  • Ben Letaifa Zouari, Leila
  • De Velasco Vázquez, Mikel
  • Justo Blanco, Raquel
  • Torres Barañano, María Inés
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Contrasting the Emotions identified in Spanish TV debates and in Human-Machine Interactions

  • De Velasco Vázquez, Mikel
  • Justo Blanco, Raquel
  • Ben Letaifa Zouari, Leila
  • Torres Barañano, María Inés
Conferencia presentada en IberSPEECH 2021, 24-25 March 2021, Valladolid, Spain, This work is aimed to contrast the similarities and differences for the emotions identified in two very different scenarios: human-to-human interaction on Spanish TV debates and human-machine interaction with a virtual agent in Spanish. To this end we developed a crowd annotation procedure to label the speech signal in terms of both, emotional categories and Valence-Arousal-Dominance models. The analysis of these data showed interesting findings that allowed to profile both the speakers and the task. Then, Convolutional Neural Networks were used for the automatic classification of the emotional samples in both tasks. Experimental results drew up a different human behavior in both tasks and outlined different speaker profiles., The research presented in this paper is conducted as part of the AMIC and EMPATHIC projects project that have received funding from the Spanish Minister of Science under grant TIN2017-85854-C4-3-R and from the European Union’s Horizon 2020 research and innovation program under grant agreement No 769872. First author has also received a PhD scholarship from the University of the Basque Country UPV/EHU, PIF17/310.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

A Spanish Corpus for Talking to the Elderly

  • Justo Blanco, Raquel
  • Ben Letaifa Zouari, Leila
  • Olaso Fernández, Javier Mikel
  • López Zorrilla, Asier
  • De Velasco Vázquez, Mikel
  • Vázquez Risco, Alain
  • Torres Barañano, María Inés
Paper presented at 11th International Workshop on Spoken Dialogue Systems, IWSDS 2020; Madrid; Spain; 21 September 2020 through 23 September 2020, In this work, a Spanish corpus that was developed, within the EMPATHIC project (http://www.empathic-project.eu/) framework, is presented. It was designed for building a dialogue system capable of talking to elderly people and promoting healthy habits, through a coaching model. The corpus, that comprises audio, video an text channels, was acquired by using a Wizard of Oz strategy. It was annotated in terms of different labels according to the different models that are needed in a dialogue system, including an emotion based annotation that will be used to generate empathetic system reactions. The annotation at different levels along with the employed procedure are described and analysed.
Addi. Archivo Digital para la Docencia y la Investigación
Publicación Contribución de congreso (ConferenceOutput)

Transfer learning in hierarchical dialogue topic classification with neural networks

  • Montenegro Portillo, César
  • Santana Hermida, Roberto
  • Lozano Alonso, José Antonio
Knowledge transfer between tasks can significantly improve the efficiency of machine learning algorithms. In supervised natural language understanding problems, this sort of improvement is critical since the availability of labelled data is usually scarce. In this paper we address the question of transfer learning between related topic classification tasks. A characteristic of our problem is that the tasks have a hierarchical relationship. Therefore, we introduce and validate how to implement the transfer exploiting this hierarchical structure. Our results for a real-world topic classification task show that the transfer can produce improvements in the behavior of the classifiers for some particular problems., The research presented in this paper is conducted as part of the project EM-PATHIC that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 769872.