Universidad del País Vasco
Universidad del País Vasco
The University of the Basque Country (UPV/EHU) is a teaching and research institution officially founded in 1985, that is in the top five Spanish Universities. The university employs over 7.000 people in 31 faculties and schools distributed in three different campuses, with a total of more than 50,000 undergraduate and postgraduate students. The UPV/EHU offers the highest number of doctorate programmes of all Spanish universities, one third of which have received a mention of excellence from the Spanish Ministry of Education. The UPV/EHU has been recognised as an International Research Campus of Excellence by the Spanish Ministry of Science and Innovation.
Since the first Research Framework Programmes, the UPV/EHU has been very active and has participated in many collaborative projects and Marie Curie actions. With regards to the 7th Framework Programme, during the period 2007-2013, the University of the Basque Country has participated in 94 projects, has coordinated 22 of them and has been the beneficiary of 5 ERC Grants, amounting to more than 30 million Euros of financing from the FP7 programme, and more than 3 million Euros from other European or International Research Programmes.
In Horizon 2020 the UPV/EHU has already been successful with 42 projects which so far have attracted a total sum of 12 million Euros in research financing, including one ERC Advanced Grant and one FET-OPEN grant.
Within this project UPV/EHU will be represented by two groups, the Speech Interactive research group and the Intelligent Systems group.
Speech Interactive research group
The Speech Interactive research group that was formed in 1990 to develop research on pattern recognition and speech technologies. The group has dealt with automatic speech recognition, understanding and translation using data-driven approaches as well as in generating language resources. It has also more than ten years of experience in the development of spoken dialog systems, specifically on architectures for practical applications and on finite-state and agenda-based managers and their adaptation to new tasks. The group has also developed statistical managers based on stochastic bi-automata as well as methods to learn from human interaction for task and user adaptation. Statistical topic classifiers and statistical suggestion generator for active turn taking are also of interest of the group.
The group now also investigates the emotional aspects of interaction, focusing on the emotion detection from the analysis of speech signal as well as from the dialogic language in social media. Research on emotion detection from speech signal is being carrying out on a telephone based customer assistance domain. It focuses on tracking changes in personal acoustic, prosodic and lexical parameters associated to mood changes. Research on emotion detection from language is devoted to select statistical, semantic and syntactic features as well as context knowledge to automated identification of secondary emotions, such as sarcasm, in dialogic language in social web.
The group has been very active in leading National projects related to speech and language technologies funded by the Spanish agency of Science. It has also an important experience in research under contract with agreements with important national and some international companies. The group has now an active cooperation with the University of California (Santa Cruz) in the creation of a Spanish corpus with similar emotional annotations as the American Internet Argument Corpus (IAC) in order to carry out comparative research in both languages and cultures. (http://www.ehu.es/speech-interactive). It also collaborates with a local company in automatic detection of emotional parameters from speech in call-centres.
The group is already collaborating with other partners (IV, SUN and IMT) and other European groups in the design and development of an audio-visual corpus of dialogs for emotional lifeline human-machine interaction for elderly. This work continues the one carried out during the collaborative workshop eNTERFACE2016.
Intelligent Systems group
The Intelligent Systems group’s main research interests are intelligent data analysis (machine learning, data mining, probabilistic graphical models), optimisation (evolutionary computation, estimation of distribution algorithms) and high-performance computing (parallel implementation of algorithms, grid/cloud computing). The group is composed of 7 faculty members, 2 research fellows and 10 Ph.D. students. Their members have published more than 200 peer-reviewed papers and have organised several international conferences.
The group has participated in several projects at the European and national level. Particularly relevant is its participation as a beneficiary in the project Nature Inspire Computation and Its Application (NICaiA) Marie-Curie IRSES FP-VII. The goal of the project was to develop computational methods inspired in nature.
Particularly, most of these methods were related with the optimization field and particularly with Evolutionary Computation. At the national level, the group has kept continuous funding during the last 15 years with proposals such as: “Probabilistic modelling in machine learning and optimization: model learning, permutations and time series” or “Probabilistic Graphical Models in Machine Learning and Optimization: Efficient Implementations and Applications”.
Finally, the group has maintained a continuous and excellent relationship with companies and public administrations. Contracts have been signed with Sher.pa, Mutualia, Eustat, among others.