RL-Duet: Online Music Accompaniment Generation Using Deep Reinforcement Learning
This work addresses the challenge of interactive music generation for musicians and AI applications, though it is incremental as it builds on existing reinforcement learning and music generation techniques.
The paper tackles the problem of generating online music accompaniment for real-time human-machine duet improvisation by using deep reinforcement learning, and it shows that the proposed algorithm produces higher quality music than baseline methods in subjective evaluations.
This paper presents a deep reinforcement learning algorithm for online accompaniment generation, with potential for real-time interactive human-machine duet improvisation. Different from offline music generation and harmonization, online music accompaniment requires the algorithm to respond to human input and generate the machine counterpart in a sequential order. We cast this as a reinforcement learning problem, where the generation agent learns a policy to generate a musical note (action) based on previously generated context (state). The key of this algorithm is the well-functioning reward model. Instead of defining it using music composition rules, we learn this model from monophonic and polyphonic training data. This model considers the compatibility of the machine-generated note with both the machine-generated context and the human-generated context. Experiments show that this algorithm is able to respond to the human part and generate a melodic, harmonic and diverse machine part. Subjective evaluations on preferences show that the proposed algorithm generates music pieces of higher quality than the baseline method.