Revisiting Adversarial Autoencoder for Unsupervised Word Translation with Cycle Consistency and Improved Training
This work addresses the challenge of learning bilingual dictionaries without parallel data, offering improved robustness and results for various language pairs, though it is incremental as it builds on existing adversarial approaches.
The paper tackled the problem of unsupervised word translation by revisiting adversarial autoencoders, introducing cycle consistency and reconstruction regularization to stabilize training, and achieved better performance across European, non-European, and low-resource languages compared to recent adversarial and non-adversarial methods.
Adversarial training has shown impressive success in learning bilingual dictionary without any parallel data by mapping monolingual embeddings to a shared space. However, recent work has shown superior performance for non-adversarial methods in more challenging language pairs. In this work, we revisit adversarial autoencoder for unsupervised word translation and propose two novel extensions to it that yield more stable training and improved results. Our method includes regularization terms to enforce cycle consistency and input reconstruction, and puts the target encoders as an adversary against the corresponding discriminator. Extensive experimentations with European, non-European and low-resource languages show that our method is more robust and achieves better performance than recently proposed adversarial and non-adversarial approaches.