TASS 2018: The Strength of Deep Learning in Language Understanding Tasks

  title={TASS 2018: The Strength of Deep Learning in Language Understanding Tasks},
  author={Manuel Carlos D{\'i}az-Galiano and Miguel {\'A}ngel Garc{\'i}a Cumbreras and Manuel Garc{\'i}a Vega and Yoan Guti{\'e}rrez-V{\'a}zquez and Eugenio Mart{\'i}nez-C{\'a}mara and Alejandro Piad-Morffis and Julio Villena-Rom{\'a}n},
  journal={Proces. del Leng. Natural},
The edition of TASS in 2018 was the edition of the evolution of TASS to a competitive evaluation workshop on semantic and text understanding tasks. Consequently, TASS has enlarged the coverage of tasks, and it goes beyond sentiment analysis. Thereby, two new tasks focused on semantic relation extraction in the health domain and emotion classification in the news domain were added to the two traditional tasks of TASS, namely sentiment analysis at tweet level and aspect level. Several systems… 

Tables from this paper

Overview of TASS 2019: One More Further for the Global Spanish Sentiment Analysis Corpus
This paper summarizes the approaches and the results of the submitted systems of the different groups for each task in the TASS workshop, and proposes a new approach to sentiment analysis at tweet level.
Deep Learning Hyper-parameter Tuning for Sentiment Analysis in Twitter based on Evolutionary Algorithms
This work proposes the use of the evolutionary algorithm SHADE for the optimisation of the configuration of a deep learning model for the task of sentiment analysis in Twitter, and shows that the hyper-parameters found by the evolutionary algorithms enhance the performance of the deep learning method.
SPLORT: Evaluation of a Web-based Approach for Descriptive and Exploratory Analysis of Multivariate Spatio-temporal Data
The findings from the pilot study held suggest that SPLORT successfully allows the users to visually query, explore and analyze contextless multi-scale attributes within space and time, which quickly drives them to new insights and to support easier decision-making processes.
Del data-driven al data-feeling: análisis de sentimiento en tiempo real de mensajes en español sobre divulgación científica usando técnicas de aprendizaje automático
espanolLos cambios producidos en los ultimos anos en cuanto a modelos de comunicacion social han llevado a todos los sectores a adaptarse a los nuevos medios para alcanzar a su publico. La
Collaborative Strategy for Software Engineering Courses at a South American University
It is suggested that an intercurricular strategy like the one presented can foster students’ performance, engagement, and motivation and supports the promotion of SE skills, such as communication and teamwork.
An Augmented Reality Implementation of the Pong Game
The development of an interactive, augmented reality, and collaborative Pong game is presented, which includes natural language processing on interviews with children and qualitative analysis of their verbal responses.


Atalaya at TASS 2018: Sentiment Analysis with Tweet Embeddings and Data Augmentation
This work presents the participation as team Atalaya in the task of polarity classification of tweets, which followed standard techniques in preprocessing, representation and classification, and also explored some novel ideas.
MeaningCloud at TASS 2018: News Headlines Categorization for Brand Safety Assessment
This paper describes the participation of MeaningCloud in Task 4 at TASS 2018 (Martínez-Cámara et al., 2018), which is focused on Brand Safety assessment. The objective of systems is to predict
RETUYT-InCo at TASS 2018: Sentiment Analysis in Spanish Variants using Neural Networks and SVM
This paper presents three approaches for classifying the sentiment of tweets for different Spanish variants in the TASS 2018 challenge, based on Support Vector Machines, Convolutional Neural Netowrks and Long Short Term Memory networks.
SCI2S at TASS 2018: Emotion Classification with Recurrent Neural Networks
The participation of the team SCI2S in all the Subtasks of the Task 4 of TASS 2018 is described and three Deep Learning models that are based on a sequence encoding layer built on a Long Short-Term Memory gated-architecture of Recurrent Neural Network are proposed.
LABDA at TASS-2018 Task 3: Convolutional Neural Networks for Relation Classification in Spanish eHealth documents
The participation of the LABDA team at the subtask of classification of relationships between two identified entities in electronic health documents written in Spanish using a Convolutional Neural Network with the word embedding and the position embedding to classify the type of the relation between two entities in the sentence.
ELiRF-UPV en TASS 2018: Análisis de Sentimientos en Twitter basado en Aprendizaje Profundo (ELiRF-UPV at TASS 2018: Sentiment Analysis in Twitter based on Deep Learning)
This paper describes the participation of the ELiRF research group of the Universitat Politècnica de València at TASS2018 Workshop which is a satellite event of the XXXIV edition of the International
ELiRF-UPV en TASS 2018: Categorización Emocional de Noticias(ELiRF-UPV at TASS 2018: Emotional Categorization of News Articles)
The approaches used for “Good Or Bad News? Emotional categorization of news articles” task, the results obtained and a discussion of these results are described.
Aplicación de un modelo híbrido de aprendizaje profundo para el Análisis de Sentimiento en Twitter(Application of a hybrid deep learning model for Sentiment Analysis in Twitter)
This paper describes the participation of ITAINNOVA at sentiment analysis at Tweet level task within TASS 2018 workshop and a possible future work line which combines this architecture with the algorithm presented in the previous TASS edition.
A Hybrid Bi-LSTM-CRF model for Knowledge Recognition from eHealth documents
A Deep Learning architecture for Named Entity Recognition (NER) in biomedical texts has two bidirectional Long Short-Term Memory layers and a last layer based on Conditional Random Field.
INGEOTEC solution for Task 1 in TASS'18 competition
This paper describes the participating system in TASS’18, composed of several subsystems independently collected and trained, combined with the EvoMSA genetic programming system.