Adaptive Compression of the Latent Space in Variational Autoencoders (2312.06280v1)
Abstract: Variational Autoencoders (VAEs) are powerful generative models that have been widely used in various fields, including image and text generation. However, one of the known challenges in using VAEs is the model's sensitivity to its hyperparameters, such as the latent space size. This paper presents a simple extension of VAEs for automatically determining the optimal latent space size during the training process by gradually decreasing the latent size through neuron removal and observing the model performance. The proposed method is compared to traditional hyperparameter grid search and is shown to be significantly faster while still achieving the best optimal dimensionality on four image datasets. Furthermore, we show that the final performance of our method is comparable to training on the optimal latent size from scratch, and might thus serve as a convenient substitute.
- Examining the size of the latent space of convolutional variational autoencoders trained with spectral topographic maps of eeg frequency bands. IEEE Access, 10:107575–107586, 2022.
- Diagnosing and enhancing vae models. arXiv preprint arXiv:1903.05789, 2019a.
- Diagnosing and enhancing VAE models. In 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, 2019b.
- Li Deng. The mnist database of handwritten digit images for machine learning research. IEEE Signal Processing Magazine, 29(6):141–142, 2012.
- Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019.
- Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30, 2017.
- beta-vae: Learning basic visual concepts with a constrained variational framework. 2016.
- Parameter tuning is a key part of dimensionality reduction via deep variational autoencoders for single cell rna transcriptomics. In BIOCOMPUTING 2019: Proceedings of the Pacific Symposium, pp. 362–373. World Scientific, 2018.
- Disentangling by factorising. In International Conference on Machine Learning, pp. 2649–2658. PMLR, 2018.
- Auto-Encoding Variational Bayes. In 2nd International Conference on Learning Representations (ICLR), 2014.
- Disentangled sequential autoencoder. In International Conference on Machine Learning, 2018.
- Challenging common assumptions in the unsupervised learning of disentangled representations. In international conference on machine learning, pp. 4114–4124. PMLR, 2019.
- Maskaae: Latent space optimization for adversarial auto-encoders. In Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), volume 124, pp. 689–698. PMLR, 2020.
- Sample complexity of testing the manifold hypothesis. Advances in neural information processing systems, 23, 2010.
- Controlvae: Controllable variational autoencoder. In International Conference on Machine Learning, pp. 8655–8664. PMLR, 2020.
- Compressing gene expression data using multiple latent space dimensionalities learns complementary biological representations. Genome biology, 21(1):1–27, 2020.
- Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms, 2017.
- Videogpt: Video generation using vq-vae and transformers. arXiv preprint arXiv:2104.10157, 2021.