VRAIN develops natural language processing and multimedia AI technologies for speech recognition, machine translation, sentiment analysis, and multimodal content understanding.
Language technologies developed at VRAIN include the TransLectures-UPV platform for multilingual transcription and translation of video content, TT-Streaming for live speech recognition and real-time translation of audio streams, and neural models covering Spanish, Catalan/Valencian, English, German, Portuguese, Italian, Dutch, French, and Slovenian. Further tools include SENTAT for sentiment analysis of Spanish-language tweets, EN-IRONIC and ES-IRONIC for irony detection in English and Spanish tweets respectively, and systems for text simplification and automatic subtitling.
Applications extend to recommendation systems for multimedia content, named entity recognition for information extraction, and affective computing for understanding emotional content in audiovisual media. Sectors served include education, parliamentary proceedings, legal document processing, and media production.
Related Technological Offer
Software & Patents
- Communication — TransLectures-UPV multilingual platform, TT-Streaming real-time translation, SENTAT sentiment analysis, speech synthesis and recognition systems, automatic translation, SPA-Sentences, EN-IRONIC and ES-IRONIC irony detection, pyTLK speech toolkit, Software TTT.
Services
- Communication — Simplification of texts for improved accessibility.
- Reasoning — Detection of named entities, recommendation systems for content.
- Planning — Monitoring of social networks for sentiment and trend analysis.
Related Research Areas
- Natural Language Processing — Speech recognition, machine translation, sentiment analysis, text simplification.
- Machine Learning and Deep Learning — Multimodal content analysis, audiovisual AI.
- Intelligent Agents and Human-Centered AI — Affective computing, recommendation systems.