VoiceSpeech technologies, language consultancy and solutions for vocal interfaces
With respects to human-machine relationship based on verbal interaction – Siri di Apple being the most recognised – the application of NLP technology and the use of specialised skills in the elaboration of language are elements of value in the realisation of high quality voice synthesis system. Relying on a language technology capable of extracting speech-related information (i.e. prosodic and phonetic information) from written texts is essential to synthesize a true natural speech.
Each language requires a specific language and phonetic study, for this reason we develop specialized versions of our software components, according to a modular approach (specific language modules).
- Study of languages orientated to the needs of vocal synthesis – phonetical aspects, prosodic, morphological and synthetic
- Consultancy on various languages, native-speaker linguists with elevated levels of specialisation
- Creation of phonetical grammar, lexicons, written texts and other personalised linguistic resources
Natural Language Interfaces
Conversational systems are designed to provide users with a simple and natural interaction with systems, using the tool they know best: their language. “Chatbots”, “spoken dialogue systems”, “conversational computing”, “conversational agents”, “Natural Human-Computer Interaction” are some of the terms used talking about interactive frameworks integrating human language.
In specific industries or applications, such as eg. e-commerce, banking and insurance, health and personal care, enterprise help desk, public administration users have different and precise needs. In these cases the natural dialogue system needs to lean on a detailed knowledge of the domain. At the same time it must be able to understand complex interactions, in order to answer properly and satisfying the users’ needs, trying to understand their intentions.
Natural dialogue systems therefore have to face a twofold challenge. Firstly, we have to face the complexity of human language, in order to understand it. Natural Language Processing and Machine Learning offer good solutions to this issue. Secondly, domain knowledge has to be used by the dialogue system in order to turn natural interactions into value.
CELI integrates its Natural Language Processing and Machine learning technologies in dialogue systems and natural interfaces. We analyse linguistic data provided by users in depth, in order to design a great user experience in natural interaction. Natural Language Understanding (NLU) and Generation (NLG), and Dialogue management, are at the center of the system. They integrate with Automatic Speech Recognition (ASR) and Text To Speech (TTS), components; the Dialogue Manager interacts with the knowledge base of the specific application. The architecture allows to design a positive user experience, and at the same to implement an effective interaction with systems in a short time.