Category Research

On 7-11 May, AIM researchers will participate at the Twelfth International Conference on Learning Representations (ICLR 2024), taking place in Vienna, Austria. ICLR is the premier gathering of professionals dedicated to the advancement of the branch of artificial intelligence called representation learning, but generally referred to as deep learning.

AIM members will be presenting the following paper at the main track of ICLR 2024:

• MERT: acoustic music understanding model with large-scale self-supervised training, by Yizhi Li, Ruibin Yuan, Ge Zhang, Yinghao MA, Xingran Chen, Hanzhi Yin, Chenghao Xiao, Chenghua Lin, Anton Ragni, Emmanouil Benetos, Norbert Gyenge, Roger Dannenberg, Ruibo Liu, Wenhu Chen, Gus Xia, Yemin Shi, Wenhao Huang, Zili Wang, Yike Guo, Jie Fu (postprint)

AIM members will also be presenting the following paper at the ICLR 2024 workshop on LLM Agents:

• WavCraft: Audio Editing and Generation with Large Language Models, by Jinhua Liang, Huan Zhang, Haohe Liu, Yin Cao, Qiuqiang Kong, Xubo Liu, Wenwu Wang, Mark D Plumbley, Huy Phan, Emmanouil Benetos (postprint)

See you all at ICLR!

On 14-19 April, several AIM researchers will participate at the 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2024). ICASSP is the leading conference in the field of signal processing and the flagship event of the IEEE Signal Processing Society.As in previous years, the UKRI Centre for Doctoral Training in AI and Music will have a strong presence at the conference, both in terms of numbers and overall impact. The below papers presented at ICASSP 2024 are authored or co-authored by AIM students:

• High resolution guitar transcription via domain adaptation, by Xavier Riley, Drew Edwards, and Simon Dixon
• Bass accompaniment generation via latent diffusion, by Marco Pasini, Maarten Grachten, and Stefan Lattner
• Unsupervised pitch-timbre disentanglement of musical instruments using a Jacobian disentangled sequential autoencoder, by Yin-Jyun Luo, Sebastian Ewert, and Simon Dixon
• Posterior variance-parameterised Gaussian dropout: improving disentangled sequential autoencoders for zero-shot voice conversion, by Yin-Jyun Luo and Simon Dixon
• Mertech: instrument playing technique detection using self-supervised pretrained model with multi-task finetuning, by Dichucheng Li, Yinghao Ma, Weixing Wei, Qiuqiang Kong, Yulun Wu, Mingjin Che, Fan Xia, Emmanouil Benetos, and Wei Li
• Syncfusion: Multimodal Onset-Synchronized Video-to-Audio Foley Synthesis, by Marco Comunità, Riccardo F. Gramaccioni, Emilian Postolache, Emanuele Rodolà, Danilo Comminiello, and Joshua D. Reiss

The following paper will be presented at the ICASSP Workshop on Explainable Machine Learning for Speech and Audio:

• Explainable modeling of gender-targeting practices in toy advertising sound and music, by Luca Marinelli and Charalampos Saitis

See you all at ICASSP!