CONF Kumar_NEURIPS2025_2025/IDIAP Latent Space Factorization in LoRA Kumar, Shashi Kaloga, Yacouba Mitros, John Motlicek, Petr Kodrasi, Ina fvae-lora latent space factorization LoRA low-rank adaptation spurious correlation robustness EXTERNAL http://publications.idiap.ch/attachments/papers/2025/Kumar_NEURIPS2025_2025.pdf PUBLIC 39th Conference on Neural Information Processing Systems 2025 https://arxiv.org/abs/2510.19640 URL Low-rank adaptation (LoRA) is a widely used method for parameter-efficient finetuning. However, existing LoRA variants lack mechanisms to explicitly disambiguate task-relevant information within the learned low-rank subspace, potentially limiting downstream performance. We propose Factorized Variational Autoencoder LoRA (FVAE-LoRA), which leverages a VAE to learn two distinct latent spaces. Our novel Evidence Lower Bound formulation explicitly promotes factorization between the latent spaces, dedicating one latent space to task-salient features and the other to residual information. Extensive experiments on text, audio, and image tasks demonstrate that FVAE-LoRA consistently outperforms standard LoRA. Moreover, spurious correlation evaluations confirm that FVAE-LoRA better isolates task-relevant signals, leading to improved robustness under distribution shifts. Our code is publicly available at: \url{https://github.com/idiap/FVAE-LoRA}