US 11,705,096 B2
Autonomous generation of melody
Shaohan Huang, Redmond, WA (US); Lei Cui, Beijing (CN); Tao Ge, Redmond, WA (US); Furu Wei, Redmond, WA (US); and Ming Zhou, Beijing (CN)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Appl. No. 17/51,409
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
PCT Filed May 24, 2019, PCT No. PCT/US2019/033866
§ 371(c)(1), (2) Date Oct. 28, 2020,
PCT Pub. No. WO2019/231835, PCT Pub. Date Dec. 5, 2019.
Claims priority of application No. 201810558917.1 (CN), filed on Jun. 1, 2018.
Prior Publication US 2021/0158790 A1, May 27, 2021
Int. Cl. G10H 1/00 (2006.01); G06N 20/00 (2019.01)
CPC G10H 1/0025 (2013.01) [G06N 20/00 (2019.01); G10H 2210/111 (2013.01); G10H 2220/351 (2013.01); G10H 2240/081 (2013.01); G10H 2240/085 (2013.01); G10H 2250/311 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
detecting a user emotion and/or environmental information;
selecting a first melody feature parameter corresponding to a first reference melody from a plurality of melody feature parameters based on the user emotion and/or the environmental information, each of the plurality of melody feature parameters corresponding to a respective one of a plurality of reference melodies, wherein each reference melody is encoded via an encoder of a variational autoencoder (VAE) trained on the reference melodies to generate the plurality of melody, feature parameters; and
generating, via a decoder of the VAE, a first melody conforming to the music style based on the first melody feature parameter; the first melody being different from the first reference melody.