TALP RESEARCH CENTER

Mission:

To provide new technological applications for oral and written language processing that revert the benefits to society.

 


Vision:

To be a reference center in innovation and development of language technologies and in the construction of systems capable of processing human language; facilitating communication between people; interacting with people through devices; and accessing to multimodal information .

The management of the TALP research center expresses its commitment to the promotion of continuous improvement in the management of research, development and the transfer of language and speech technologies, taking into account:

  • The principles and values ​​that govern the ethical code and good practices of the personnel at the service of the UPC and of their students in the exercise of their activities.
  • The design of mechanisms, policies and guidelines to achieve a proven and competitive scientific quality.
  • Participation and cooperation with different agents of the society maintaining independence, prestige and social projection.
  • Optimization in the use of resources with criteria of efficiency, efficiency, austerity and sustainability to manage the activities and development of people and work groups with impartiality and equity.
  • Commitment to the client and to the Safety and Health of the people.

 


Ethics

The research center pursues rigorous application of ethical principles in all areas of application of its research. The issues in which we may be involved are related to:

  • Involvement of children and patients in medical applications. One of the lines of work is related to the detection of problems in neonatal rooms. We have experience in collecting data to perform this task and it should be mentioned that both, the VEU group and the hospitals we have worked with, are extraordinarily meticulous with ethical and data privacy issues. It is true that these restrictions shorten the number of hours of recording needed to train automatic learning systems, but we have always been and will continue to be extremely careful in this project.
  • The privacy and confidentiality of personal data in obtaining oral or multimodal databases to train or test voice recognition and synthesis systems. Any recording requires, according to our legislation, a signed document (or approval) by the person who gives his voice on a recording. Our models of acceptance of recordings participation in a research project and confidentiality clauses have been verified by our legal services and we have disseminated them as a model to other research groups. We will continue in this project in the same line of confidentiality, privacy and data protection.
  • We have also signed confidential agreements when using databases to train translation models. We have been entrusted with these databases worth over 2 million euros and we have used special protection strategies in our servers to ensure its protection.
  • Plagiarism or any other inappropriate behavior.
Scroll to Top