Complexity of symbolic representation in working memory of Transformer correlates with the complexity of a task |
| |
Institution: | 1. Ruhr-University Bochum, Institute for Philosophy II, Bochum, Germany;2. The University of Edinburgh, School of Philosophy, Psychology and Language Sciences, United Kingdom;3. Amsterdam University Medical Centre, Department of Psychiatry, Netherlands;4. Amsterdam Brain and Cognition, Netherlands |
| |
Abstract: | 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. |
| |
Keywords: | Neuro-symbolic representation Transformer Working memory Machine translation |
本文献已被 ScienceDirect 等数据库收录! |
|