Conference paper Open Access

MINGUS: Melodic Improvisation Neural Generator Using Seq2Seq

Vincenzo Madaghiele; Pasquale Lisena; Raphael Troncy

JSON-LD ( Export

  "description": "Sequence to Sequence (Seq2Seq) approaches have shown good performances in automatic music generation. We introduce MINGUS, a Transformer-based Seq2Seq architecture for modelling and generating monophonic jazz melodic lines.\n MINGUS relies on two dedicated embedding models (respectively for pitch and duration) and exploits in prediction features such as chords (current and following), bass line, position inside the measure. \n The obtained results are comparable with the state of the art of music generation with neural models, with particularly good performances on jazz music.", 
  "license": "", 
  "creator": [
      "@type": "Person", 
      "name": "Vincenzo Madaghiele"
      "@type": "Person", 
      "name": "Pasquale Lisena"
      "@type": "Person", 
      "name": "Raphael Troncy"
  "headline": "MINGUS: Melodic Improvisation Neural Generator Using Seq2Seq", 
  "image": "", 
  "datePublished": "2021-11-07", 
  "url": "", 
  "@type": "ScholarlyArticle", 
  "@context": "", 
  "identifier": "", 
  "@id": "", 
  "workFeatured": {
    "url": "", 
    "alternateName": "ISMIR 2021", 
    "location": "Online", 
    "@type": "Event", 
    "name": "International Society for Music Information Retrieval Conference"
  "name": "MINGUS: Melodic Improvisation Neural Generator Using Seq2Seq"
All versions This version
Views 9797
Downloads 6868
Data volume 45.1 MB45.1 MB
Unique views 9090
Unique downloads 5959


Cite as