A5101466477- Speech and dialogue systems
- Topic Modeling
- Natural Language Processing Techniques
- Subtitles and Audiovisual Media
- Multi-Agent Systems and Negotiation
- Translation Studies and Practices
- Speech Recognition and Synthesis
- Voice and Speech Disorders
- AI in Service Interactions
- Music and Audio Processing
- Multimedia Communication and Technology
- Phonetics and Phonology Research
- Social Robot Interaction and HRI
- Linguistic Studies and Language Acquisition
- Video Analysis and Summarization
- Radio, Podcasts, and Digital Media
- Speech and Audio Processing
- Technology Use by Older Adults
- Software Engineering Research
- Multimodal Machine Learning Applications
- Dysphagia Assessment and Management
- Innovation Policy and R&D
- Text Readability and Simplification
- Business, Education, Mathematics Research
- Mobile Health and mHealth Applications
Vicomtech
2014-2024
Digital Research Alliance of Canada
2022
Tech Foundation
2022
University of Zurich
2012-2018
University of Cambridge
2006-2008
Bridge University
2006
This paper presents a question answering (QA) system that will enable workers from the manufacturing industry to 'hands-free' request information. kind of systems, are broadly used in household context, have started gain popularity industrial environments. To develop system, PDF-based manuals Spanish been processed, annotated by experts and train different components architecture, i.e. classifier two QA systems. Different metrics applied developed order test performance system. The...
Most Voice Conversion (VC) systems exploit source-filter decomposition based on linear prediction (LP) to transform spectral envelopes, incurring as a result various issues related the oversimplification of LP voice source model.Whilst residual methods can mitigate this problem, they cannot be used modify quality.In paper, system which employs transformations convert both envelope and LF glottal waveform is presented.Its performance shown comparable that state-of-theart VC implementation in...
This paper describes an investigation into the repair of continuous tracheoesophageal (TE) speech. Our system resynthesises TE speech using a synthetic glottal waveform, reduces its jitter and shimmer applies novel spectral smoothing tilt correction algorithm, derived from comparative study normal envelopes. The perceptual enhancement achieved by each performance whole are evaluated on corpus thirteen speakers listening tests. Results show that our algorithms reduce perceived breathiness...
Automatic subtitling of television content has become an approachable challenge due to the advancement technology involved. In addition, it also a priority need for many Spanish TV broadcasters, who will have broadcast up 90% subtitled by 2013 comply with recently approved national audiovisual policies. APyCA, prototype system described in this paper, been developed attempt automate process through application state-of-the-art speech and language technologies. Voice activity detection,...
The population in the world is aging dramatically, and therefore, economic social effort required to maintain quality of life being increased. Assistive technologies are progressively expanding present great opportunities; however, given sensitivity health issues vulnerability older adults, some considerations need be considered. This paper presents DigiHEALTH, a suite digital solutions for long-term healthy active aging. It result fruitful trajectory research where we have understood...
Technology roadmapping provides a strategic tool to help companies develop an outside‐in view and challenge their current competitive perspectives. In this paper, the authors describe process, which is aligned, with research development ( R & D ) strategy of applied centre. This process based in adapted combination state‐of‐the‐art methodologies, as result, case study shows interesting findings terms strategy, technology processes methodology.
This paper describes an investigation into the repair of prosodic limitations tracheoesophageal (TE) speech.The proposed algorithm modifies TE phone durations based on predictions regression trees built from nonpathological data.Acoustic and language modelling refinements for improved recognition, studies feature relevance duration prediction a robust modification method are also presented.Objective subjective evaluation results show that pattern repaired sentences is closer to normal...
The aim of the AnHitz project, whose participants are research groups with very different backgrounds, is to carry out on language, speech and visual technologies for Basque. Several resources, tools applications have been developed in AnHitz, but we also integrated many these into a prototype 3D virtual expert science technology. It includes Question Answering Cross Lingual Information Retrieval systems those areas. interaction system carried Basque (the results CLIR module that not...