 Continuem ara amb la Mónica Domínguez, amb el seu treball titulat Speech Prosody in Human Machine Interaction. Hi. I think I've got... yeah. I'd like to propose a game. You can choose here two sentences, one by me and one by my friend here, the synthesizer. Sentences number one. You know what I mean. Sentences number two. Okay. So, differences. Intensity. Yeah, I'm much louder, I know. But I was also stressing, you know what I mean. And the synthesizer was like, you know what I mean. Okay, intensity. Duration, I was like, you know what I mean. And the synthesizer was regular, you know what I mean. And intonation, you know what I mean. And synthesizer, you know what I mean. Okay. Three acoustic features that characterize Speech Prosody. And you will agree with me that, well, Speech Prosody is a language and we communicate, what we communicate is not only what we say, but also how we say it. So you may wonder why does your friend speak so weirdly. Well, it's basically the way we teach computers. We feed computers with a lot of examples. We record human voices and we label and segment those data. And, for example, we say, you know what to do. And we label that as subject-object, sorry, subject verb object. If you didn't do computer, computer learns, and when we tell it. And now you need to read, you know what I mean. Because the database finds you know what to do and emphasis on what. Wrong. What am I doing to solve this problem? Well, I'm looking into a linguistic theory called information structure. That says content is packed into communicative units. And what I'm trying to look is how these communicative units are related to prosody. So, again, more game, find the differences. What's the difference? Okay, in speech technologies, we start from text and generate speech. Well, what I'm trying to do is start from content and generate speech. And when we look at content, we do care about communication. And when we care about communication, human-machine interaction can improve. In this scenario, you, yeah, research on prosody on the one hand, research on linguistics on the other hand, together with computer science, we do have a lot to say. And it's not only what we say, but how we say it. You know what I mean. Thank you.