7 de August de 2016
FrameNet Brazil develops virtual travel guide
m.knob, Multilingual Knowledge Base, is a virtual travel guide that provides personalized recommendations of tourist attractions and events. It is the result of an interdisciplinary research project conducted at the FrameNet Brasil Laboratory, with researchers and graduate students from the Linguistics and Computer Science departments of Federal University of Juiz de Fora.
m.knob combines natural language understanding with ontologies and linked data for an innovative touristic experience.
m.knob was released during the Rio Olympic Games in order to take advantage of an unique opportunity for testing our scientific hypotheses: we can hardly ever bring together to the same city people from so many different parts of the world, with diverse tourism preferences and language skills.
The app has two main functions: the Local Guide and the Dictiopedia.
In the Local Guide, users receive recommendations of tourist attractions and events based on preferences informed to the application. The recommendations are grouped into two tabs: Tourism, which shows the traditional attractions of Rio de Janeiro, and Sports, bringing the Olympic Games venues and related attractions. An additional tab shows what the user can visit nearby. In addition to the recommendations based on the profile, the user can “tell” the app want she wants to do. So, if one types “I want to eat” or “I want to watch the games,” the app will return places where she can dine or watch a game. From the list of recommendations, the user can access more detailed information about each of the recommended sites, such as map, description, hours of operation and the official website. The content information is automatically extracted from available online open databases such as Google Places and Wikidata, while the semantic processing of texts uses a multilingual lexical database locally developed by FrameNet Brasil at the Federal University of Juiz de Fora.
In the Diciopédia, the user have access to tourism and sports vocabularies in Brazilian Portuguese, English and Spanish. With this release, there will be around 2,500 words and expressions available and that number will grow every day, since users can suggest new words and definitions, as well as report translation inadequacies. The translations presented by the app are also automatically generated via two different methodologies. For verbs and names that indicate events, they are calculated by a system that compares uses of these words in semantically annotated real texts stored in the FrameNet Brasil database. For names that indicate people, objects and places, they are automatically extracted from BabelNet, an open online database, and validated by the linguists in the project.
For m.knob, user feedback is very important. Therefore, all interactions with the app are scored on a gamification platform in which a character – Greg, the “gringo” – wins badges and trophies related to user’s achievements. For every new achievement, new clothes are unlocked and Greg becomes less of a “Turistão” and more like “De Casa”.
And the app development will not stop with the end of the Olympics. The sports features will replaced by a semantically grounded sentence translation system. Therefore, m.knob will continue to collaborate with the mission of FrameNet Brasil: to develop interdisciplinary research on Computational Linguistics applied to the needs of society.
Main researchers:
– Tiago Torrent – Linguistics Department
– Ely Matos – Knowledge Management Center
– Fernanda Campos – Computer Science Department
– Regina Braga – Computer Science Department
– Margarida Salomão – Linguistics Department
– Meire Lara – Linguistics Department
– Tiago Torrent – Linguistics Department
– Ely Matos – Knowledge Management Center
– Fernanda Campos – Computer Science Department
– Regina Braga – Computer Science Department
– Margarida Salomão – Linguistics Department
– Meire Lara – Linguistics Department
Graduate Student Researchers:
– Alexandre Diniz – Linguistics
– Maucha Andrade – Linguistics
– Simone Peron-Correa – Linguistics
– Leandro Simões – Computer Science
– Adrieli Laviola – Linguistics
– Ludmila Lage – Linguistics
– Natalia Duarte – Linguistics
– Tatiane Tavares – Linguistics
– Vania Almeida – Linguistics
– Maucha Andrade – Linguistics
– Simone Peron-Correa – Linguistics
– Leandro Simões – Computer Science
– Adrieli Laviola – Linguistics
– Ludmila Lage – Linguistics
– Natalia Duarte – Linguistics
– Tatiane Tavares – Linguistics
– Vania Almeida – Linguistics
Undergraduate Students:
– Ana Carolina Alcantara – Linguistics
– Carina Silveira – Linguistics
– Elida Ramos – Linguistics
– Luke Fazola – Linguistics
– Thassya Abreu – Computer Science
– Carina Silveira – Linguistics
– Elida Ramos – Linguistics
– Luke Fazola – Linguistics
– Thassya Abreu – Computer Science
Technical staff:
– Marcelo Viridiano – User Interface Design
– Mateus Dalbianco – Webdesign
– Welington Veiga – Frontend Development
– Mateus Dalbianco – Webdesign
– Welington Veiga – Frontend Development
– Martina Fantini – Character Concept
– Lucas Nova – Trophies and badges
– Kureb – Logo Design
– Daniel Morais – Teaser Animation
– Luiz Fernando Priamo – Teaser Direction
– Ulisses Beleigoli – Teaser voice
– Gabriel Zambon – Teaser audio
Link: mknob.com
Contact: app.mknob@ufjf.edu.br