Complexity of Symbolic Representation in Working Memory of Transformer Correlates with the Complexity of a Task
This addresses the problem of memory limitations in Transformers for machine translation, offering a neural-symbolic enhancement that is incremental in nature.
The paper tackles the lack of explicit memory in Transformers for NLP tasks by adding a symbolic working memory to the decoder, which improves machine translation quality and stores relevant keywords, with memory content diversity correlating with corpus complexity.
Even though Transformers are extensively used for Natural Language Processing tasks, especially for machine translation, they lack an explicit memory to store key concepts of processed texts. This paper explores the properties of the content of symbolic working memory added to the Transformer model decoder. Such working memory enhances the quality of model predictions in machine translation task and works as a neural-symbolic representation of information that is important for the model to make correct translations. The study of memory content revealed that translated text keywords are stored in the working memory, pointing to the relevance of memory content to the processed text. Also, the diversity of tokens and parts of speech stored in memory correlates with the complexity of the corpora for machine translation task.