Deep Learning, Ian Goodfellow, Yoshua Bengio, and Aaron Courville, 2016 (MIT Press) - This foundational textbook offers comprehensive coverage of deep learning principles, including autoencoders, convolutional neural networks, and recurrent neural networks, providing a strong basis for understanding their advanced architectural variants.
Adversarial Autoencoders, Alireza Makhzani, Jonathon Shlens, Navdeep Jaitly, Ian Goodfellow, Brendan Frey, 2015arXiv preprint arXiv:1511.05644DOI: 10.48550/arXiv.1511.05644 - Introduces the Adversarial Autoencoder (AAE) architecture, which combines autoencoders with Generative Adversarial Networks (GANs) to match the aggregated posterior of the latent space to an arbitrary prior distribution.
Neural Discrete Representation Learning, Aaron van den Oord, Oriol Vinyals, Koray Kavukcuoglu, 2017Advances in Neural Information Processing Systems, Vol. 30DOI: 10.48550/arXiv.1711.00937 - Introduces the Vector Quantized Variational Autoencoder (VQ-VAE), an architecture for learning discrete latent representations that can overcome issues like posterior collapse and improve generative quality.
Attention Is All You Need, Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, Illia Polosukhin, 2017Advances in Neural Information Processing Systems, Vol. 30DOI: 10.48550/arXiv.1706.03762 - This seminal paper introduces the Transformer architecture, based entirely on attention mechanisms, which has become foundational for sequence-to-sequence tasks and is the basis for Transformer-based autoencoders.