Polifonia to present novel music segmentation method based on symbolic chord annotations at AIxIA 2022

Polifonia’s team from Bologna is gearing up for 21st International Conference of the Italian Association for Artificial Intelligence (AIxIA 2022), where Nicolas Lazzari, Andrea Poltronieri and Valentina Presutti are presenting a novel music segmentation method called Pitchclass2vec.

28 November 2022

Polifonia’s team from Bologna is gearing up for 21st International Conference of the Italian Association for Artificial Intelligence (AIxIA 2022), where Nicolas Lazzari, Andrea Poltronieri and Valentina Presutti are presenting a novel music segmentation method called Pitchclass2vec.

A new research paper Pitchclass2vec: Symbolic Music Structure Segmentation with Chord Embeddings will be presented at the CREAI 2022 First Workshop on Artificial Intelligence and Creativity. The workshop is held in conjunction with AIxIA 2022 organised by the Italian Association for Artificial Intelligence. This paper presents a new approach to automatic music segmentation based on musical chords. Traditionally, music segmentation involves the identification of the musical structures that make up a piece of music – such as Introduction, Verse, Chorus or Bridge – and is obtained from the audio signal. However, the purpose of this study was to obtain this information only from musical chord data.

The researchers trained a Deep Learning algorithm on a large dataset of songs annotated with musical chords and their musical structures. The chords were converted into a reduced dimensional space using a novel embedding method called Pitchclass2vec, which decomposes the chord into its components and reduces the whole to a numeric vector. The trained algorithm was then run on previously unseen chord sequences in order to segment them. The results of this method outperformed the state of the art for this task.

AIxIA 2022 will take place in Udine (IT) between Monday 28th November and Friday 2nd December 2022. The complete conference program is available here. Andrea Poltronieri, Valentina Presutti and masterstudent Nicolas Lazzari will present their paper on the 29th of November, at 11.15 – 12.10 during CREAI 2022’s first session called Musical perspective. The CREAI 2022 program can be found here.

Recent News

In our newest video Paul Mulholland, leader of WP5 explains how Polifonia intends to let people access or contribute to the music data Polifonia offers, in a way that fits their level of expertise, interest or physical abilities.

In WP5  - Human Interaction with Musical Heritage - the team researches and develops highly interactive…

25 January 2023

Earlier in December, Polifonia’s team from Bologna presented a novel music segmentation method called Pitchclass2vec at the 21st International Conference of the Italian Association for Artificial Intelligence (AIxIA 2022). Master student Nicolas Lazzari was asked to present the experiments that were part of this research.

Earlier in December, Polifonia’s team from Bologna presented a novel music segmentation method called…

21 December 2022

WP4 – Musical Heritage Knowledge Extraction from text- develops methods and tools for several purposes.

WP4 - Musical Heritage Knowledge Extraction from text- develops methods and tools for several purposes.…

13 December 2022

In November Polifonia reached out to a Dutch audience. Polifonia looks back fondly on the progress presented to these Dutch heritage professionals. Find the visual report here.

In November Polifonia reached out to a Dutch audience. Polifonia looks back fondly on the progress presented…

7 December 2022

Polifonia’s team from Bologna is gearing up for 21st International Conference of the Italian Association for Artificial Intelligence (AIxIA 2022), where Nicolas Lazzari, Andrea Poltronieri and Valentina Presutti are presenting a novel music segmentation method called Pitchclass2vec.

Polifonia’s team from Bologna is gearing up for 21st International Conference of the Italian Association…

28 November 2022

Annual Conference focuses on the CLARIAH digital humanities infrastructure and its use by researchers and other professionals from the heritage field. This free program offers lectures, debates, interactive presentations, workshops and poster sessions and dives into topics like AI, social media, linked data & data stories. 

This year's CLARIAH Annual Conference focuses on the CLARIAH digital humanities infrastructure and its…

21 November 2022

From Beethoven to Metallica, from Pinkpop to the Big Sing, from the beginning of the 20th century through to last week.

From Beethoven to Metallica, from Pinkpop to the Big Sing, from the beginning of the 20th century through…

16 November 2022

Relying on the results of WP2 -ontology-based knowledge graphs- the overall goal of WP3 is to devise approaches to analysing large repositories of music (tunes, songs, etc.) to identify common, meaningful patterns that are indicative of their identity, filiation or cultural association (genres, origin, etc). Work package leader James McDermott explains the ‘Mining Musical Patterns’ work of WP3 in our latest video on YouTube. 

Relying on the results of WP2 -ontology-based knowledge graphs- the overall goal of WP3 is to devise…

9 November 2022

As part of the dissemination of the project results, Polifonia’s consortium partners have the ambition to regularly present their work at symposia and academic meetings, thereby informing semantic, AI, data science, linguistic and humanities researchers of their findings with Polifonia data.

As part of the dissemination of the project results, Polifonia's consortium partners have the ambition…

31 October 2022

Work package 2 develops ontology-based knowledge graphs for representing music collections (symbolic notation/scores, code-based musical assets, annotated audio/transcriptions, metadata) and its historical, cultural, and social context. Watch work package leader Albert Meroño Peñuela explain the role of WP2 within Polifonia in our newest YouTube video.

Work package 2 develops ontology-based knowledge graphs for representing music collections (symbolic…

17 October 2022

This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement N. 101004746