Feedback is essential for learning,’ says Carles Sierra, Research Professor at the Spanish National Research Council and coordinator of the PRAISE project. The project aims at filling a gap in online learning by creating a social network for music education with tools for giving and receiving feedback.
Using PRAISE’s Music Circle platform, music students can upload recordings of their playing and receive detailed feedback from other members of the community. Advanced tools let reviewers place their comments as annotations at exactly the right place in the audio signal representation.
‘Students’ peers can say “this crescendo is very nice” or “this passage is very expressive”,’ explains Professor Sierra. ‘This timeline of structured comments and this level of granularity have been lacking in online approaches to giving feedback on music.’
A social network of learners
As PRAISE is a web-based social media platform, a single comment can spawn multiple comments. These discussion threads help to create a community of people giving and receiving feedback that becomes a social network of learners.
Aside from human feedback, PRAISE’s sophisticated tools also provide automatic feedback. Students can play a particular piece, for example, and the software will tell the student whether he or she played the right notes at the right time. Moreover, if the student submits a new recording, areas of improvement or retrogression are similarly flagged.
A teaching tool with marks generation
PRAISE is actually more than just a platform for giving and receiving feedback. Its tools also allow music teachers to create lesson plans and track their students’ progress online.
In addition, PRAISE tackles a problem faced by many Massive Open Online Courses (MOOC). Since it’s physically impossible for a single teacher to mark the thousands or tens of thousands of students who may follow an online offering, many MOOCs rely on peer assessment. This approach is unsatisfactory, however, because students may not give marks in the same way that the teacher would give them.
To solve this problem, a teacher marks just a few students while students each mark several of their peers. PRAISE then uses an algorithm to identify the student assessments that most closely resemble those made by the teacher and these are given more weight. In this way, the teacher’s approach to giving marks ripples throughout the community of students.
‘With MOOCs, marks typically reflect an average of what students think of each other,’ says Professor Sierra. ‘We instead compute a trust network that emphasises the views of the teacher.’
A crescendo of activity
PRAISE’s Music Circle platform has already been used by over 20 000 students who finished a course in jazz piano on Coursera and it is being used by about 150 music students at Goldsmiths College of the University of London and several more at ESMUC in Barcelona.
‘One of the advantages of Music Circle is zero cost. Paying a tutor is astronomically expensive,’ notes Professor Sierra.
The project’s tools have also been applied to other domains such as to English learning.
PRAISE involves researchers from Institut d’Investigació en Intel-ligèncìa Artìfìcìal in Barcelona, Goldsmiths College of the University of London, the Vrije Universiteit Brussel and Sony’s Computer Science Lab in Paris. Part of what makes the project unique and has made the collaboration so fruitful is that PRAISE’s researchers share not only a strong interest in artificial intelligence, but also a love of music. Many of PRAISE’s researchers are gifted musicians in their own right.
PRAISE received EUR 2.3 million in EU funding and ran from 2013 to 2015.
Do you want more information on this project?
Please have a look at the related videos.