Exploiting Invertible Decoders for Unsupervised Sentence Representation Learning
This work addresses a bottleneck in NLP for researchers by enhancing unsupervised sentence embeddings, though it is incremental as it builds on existing encoder-decoder frameworks.
The paper tackled the problem of underutilizing decoder parameters in unsupervised sentence representation learning by introducing invertible decoders, enabling the inverse function to act as an additional encoder. The result showed that ensembling representations from both encoder and inverse decoder improved generalization and transferability, with specific gains in benchmark tasks.
The encoder-decoder models for unsupervised sentence representation learning tend to discard the decoder after being trained on a large unlabelled corpus, since only the encoder is needed to map the input sentence into a vector representation. However, parameters learnt in the decoder also contain useful information about language. In order to utilise the decoder after learning, we present two types of decoding functions whose inverse can be easily derived without expensive inverse calculation. Therefore, the inverse of the decoding function serves as another encoder that produces sentence representations. We show that, with careful design of the decoding functions, the model learns good sentence representations, and the ensemble of the representations produced from the encoder and the inverse of the decoder demonstrate even better generalisation ability and solid transferability.