RL-Duet: Online Music Accompaniment Generation Using Deep Reinforcement Learning
Authors: Nan Jiang, Sheng Jin, Zhiyao Duan, Changshui Zhang710-718
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 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. |
| Researcher Affiliation | Academia | 1Department of Automation, Tsinghua University State Key Lab of Intelligent Technologies and Systems Institute for Artificial Intelligence, Tsinghua University (THUAI) Beijing National Research Center for Information Science and Technology (BNRist) 2Department of Electrical and Computer Engineering, University of Rochester |
| Pseudocode | No | The paper includes a figure (Figure 4) titled 'Model architecture of the generation model' but does not present any pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain any explicit statement about releasing source code or a link to a code repository for the described methodology. |
| Open Datasets | Yes | The model is trained on the Bach Chorale dataset in Music21 (Cuthbert and Ariza 2010). |
| Dataset Splits | Yes | We use chorales with four monophonic parts, in the SATB format (Soprano, Alto, Tenor, and Bass) as the training and validation datasets, with 327 and 37 chorales respectively. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used for running the experiments (e.g., GPU models, CPU types, or memory). |
| Software Dependencies | No | The paper does not provide specific version numbers for any software dependencies used in the experiments (e.g., programming language versions, library versions). |
| Experiment Setup | Yes | The discount factor γ = 0.5. The pretrained joint-modeling reward model (a) with learning rate 0.01 is used to initialize the weights of the generation model. |