AMIS (Access Multilingual Information opinionS)
AMIS is an original project concerning the second call: Human Language Understanding; Grounding Language Learning. This project acts on different data: video, audio and text. We consider the understanding process, to be the aptitude to capture the most important ideas contained in a media expressed in a foreign language, which would be compared to an equivalent document in the mother tongue of a user. In other words, the understanding will be approached by the global meaning of the content of a support and not by the meaning of each fragment of a video, audio or text. The idea of AMIS is to facilitate the comprehension of the huge amount of information available in TV shows, internet etc. One of the possibilities to reach this objective is to summarize the amount of information and then to translate it into the end-user language. Another objective of this project is to access to the underlying emotion or opinion contained in two medias. To do this, we propose to compare the opinion of two media supports, concerning the same topic, expressed in two different languages. The idea is to study the divergence and the convergence of opinions of two documents whatever their supports. Several skills are necessary to achieve this objective: video summarization, automatic speech recognition, machine translation, language modelling, sentiment-analysis, etc. Each of them, in our consortium, is treated by machine learning techniques; nevertheless human language processing is necessary for identifying the relevant opinions and for evaluating the quality of video, audio and text summarization by the end-user.
See AMIS website for more information.
The objective of TRAM is to show the feasibility of an automatic accompaniment of Arab vocal improvisation. The idea is to propose an automatic instrumental response to an Arab singer who executes a Mawwal (or Istikhbar). The originality of the project is to investigate an approach based on Machine Translation (MT) in studying the accompaniment of Arab vocal improvisation. This approach considers the mutual interaction between the singer and the instrumentalist as a question and answer: vocal sentence (question) and instrumental response (answer). In Machine translation, we need a parallel corpus composed of a source and a target language. The training process allows then to associate each phrase of the source sentence to its corresponding phrase in the target language. To deal with this project, we propose a consortium composed of experts in music and in machine translation and more generally on machine learning process. This project necessitates collecting data which will be a considerable resource for researchers and which will be provided freely to our research community. This bootstrapping project will probably help us to apply in the near future to H2020.