The Computer Composition Identification Challenge, initiated by the Conference on Sound and Music Technology (CSMT), is succeeding the CSMT Data Challenge, 2020. This new challenge is dedicated to finding improved music generation models and objective evaluation models for a number of specified music styles.
There will be two roles competing against each other in this challenge: Generator versus Judge. Participants must choose one role and complete the corresponding tasks.
Generators are supposed to develop a system1 that can automatically generate a playable violin phrase with given beginning and ending bars, tempo value, key signature, and time signature.
Judges are supposed to analyse given music phrases and produce a Human Composition
HCS) -- the probability of the input being composed by real human composers.
1: A system can consist of more than one objective model or rule sets.
The following participants will win the prizes:
• The Best Generator for the participant whose generator system obtains the highest score.
• The Best Judge for the participant whose judge system obtains the highest score.
• Creativity Award for the participant with the most creative solution, determined by the committee.
When the challenge begins, a list of violin composers will be released, where each violin composer has a specific music style.
For generator systems, 200 music phrases should be generated in the style of each composer in the list. The beginning and ending bars will not be specified until the final submission. This means the music phrases generated before the final submission can have arbitrary beginning and ends and should be able to generate with specified beginning and ending bars in the final stage.
For judge systems, participants are required to compute the probability of a given music phrase as the input being generated by generator systems, with the given composer’s name as the context. Throughout the development stage, only the composers’ own compositions (positive samples) are allowed to be used to train the judge system .
The submission will be via Microsoft CMT. The preliminary submission entry for generators is now available. Please follow the instructions below and submit your generators.
The preliminary result contains two submission deadlines. The first deadline is for generators
to submit 200 music phrases for each composer. The second deadline is set around one week later to judges for submitting the
HCS results, which are the feedback for the
generated by generators . Notice that any valid submission record in these
two preliminary submissions will be considered as completing the registration for this challenge but the
submitted content will not affect the final score.
The last few days before the final submission deadline will be the evaluation stage, when 200 pairs of
beginnings and endings will be released for each composer's pieces in the list.
In 48 hours, generator systems should complete these music phrases and submit them as their final submission. Then, these phrases will be compiled together and made available to judge systems. From this moment on, in the next 48 hours, judge systems are expected to calculate the
HCSs for these phrases and make the final submission. For
each composer, a certain number of extra phrases extracted from human-composed pieces will be mixed into the
Generator scores are measured by are the average
HCS from all
Judge performances are ranked by the sum of the following two scores:
HCSfor human composed phrases.
1-HCSfor phrases generated by generators.
For detailed requirements, please refer to Submission Requirements .
For any enquires, contact us:
, challenge coordinator, Xi'an Jiaotong-Liverpool University,