Showing papers by "Marco Furini published in 2021"

Automatic and User-Tailored Playlist Sequencing

Abstract: Social technologies have revolutionized the world of music and playlists have become the new radios. However, the production of playlists has been mainly focused on identifying the songs that the listener might like, disregarding the songs sequencing process. In this paper, we propose a highly user-oriented sequencing method. The idea is to use the user's listening history to identify a possible sequencing criterion used by the user when playing out songs. The proposed method is then evaluated with real users and results showed the importance of personalization in the playlist sequencing process. Although the study is still in its early stage, results are promising and, in the future, we plan to personalize the sequencing process even more.

Automatic Music Playlist Generation Based on Music-Programming of FM Radios

Abstract: Streaming music services are flooding their platforms with thousands of different playlists and this huge catalog is backfiring users who struggle to find the playlist that best suits their needs. In this paper, we aim to facilitate the music listening by designing a novel approach to automatically generate a playlist suited to the user's musical taste. Our approach is based on the FM radio music programming: starting from the user's favorite radio, we analyze seven days of music programming, we transform songs into vectors of audio features, we generate a music programming for a virtual day, and we transform this virtual day music programming into a real playlist when the user begins the music playout.

Automatic and Smart Content Transformation of Video Lectures

Abstract: The lockdown caused by the Covid-19 pandemic has forced many educational institutions around the world to produce video lectures in order to support their students. A popular approach was to produce video lectures with a generic layout without considering neither the type of student nor the type of device used by the student. This approach complicated the learning process (e.g., students equipped with mobile devices and limited bandwidth connections have been forced to watch video produced for large screens and infinite bandwidth availability). In this paper, we investigate if it is possible to automatically transform the original video lecture to produce smartphone suitable videos and we involve students to understand the viewing experience. We consider the video lectures available within the ONELab University video lecture catalog and we design five different heuristics based on the semantic analysis of the video lectures. The experimental evaluation considers both the quantitative and qualitative aspects and the obtained results show that it is possible to save more than 90% of the bandwidth while maintaining a viewing experience equals to the one of the original video lectures.

On using Conversational Interfaces to Improve the Accessibility of a University Campus

Abstract: Moving around a university campus may seem an activity of little importance: students are young, they walk and run without problems, with a simple glance they know where they are and where to go, they understand where a lesson is held, when a teacher is available or where to go when they are hungry. Activities that millions of students do every day. But what about students with some form of physical impairments? For them, moving around a university campus is a daily challenge. Can conversational interfaces improve the accessibility of a University campus? is the research question we address in this study. We involved students (with and without disabilities) in the development and evaluation of a skill for the Amazon Alexa platform. Results show that conversational interfaces are highly appreciated by most of the participants and confirm that such interfaces might improve the daily experience of users within a university campus.

On Using Video Lectures Data Usage to Predict University Students Dropout

Abstract: Technologies have changed many different aspects of people's life and the recent CoVid-19 pandemic proved that education is not an exception. But technologies in education go beyond the simple use of video lectures: technologies might be exploited to improve personal learning. In this paper, we focus on the dropout of studies, a global phenomenon that artificial intelligence techniques are trying to ameliorate. Here, we investigate whether data related to the consumption of video lectures might improve the students' dropout prediction. We consider first-year students enrolled in our Department and we characterize them with personal, scholastic, academic and technological features. Then, we measure the performance of three machine learning algorithms in terms of accuracy and sensitivity. The experimental evaluation shows that Random Forest and KNN perform better that Decision Tree and also shows that data related to the use of video lectures improves the prediction performance for some degree programs (reaching 73% in terms of accuracy and sensitivity). These preliminary results show that the approach is promising and worth exploring in future studies.