On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages
Sentence embeddings are effective input vectors for the neural learning of a number of inferences about content and meaning. Unfortunately, most of such decision processes are epistemologically opaque as for the limited interpretability of the acquired neural models based on the involved embeddings....
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Accademia University Press
2019-06-01
|
| Series: | IJCoL |
| Online Access: | https://journals.openedition.org/ijcol/453 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850262882588033024 |
|---|---|
| author | Daniele Rossini Danilo Croce Roberto Basili |
| author_facet | Daniele Rossini Danilo Croce Roberto Basili |
| author_sort | Daniele Rossini |
| collection | DOAJ |
| description | Sentence embeddings are effective input vectors for the neural learning of a number of inferences about content and meaning. Unfortunately, most of such decision processes are epistemologically opaque as for the limited interpretability of the acquired neural models based on the involved embeddings. In this paper, we concentrate on the readability of neural models, discussing an embedding technique (the Nyström methodology) that corresponds to the reconstruction of a sentence in a kernel space, capturing grammatical and lexical semantic information. From this method, we build a Kernel-based Deep Architecture that is characterized by inherently high interpretability properties, as the proposed embedding is derived from examples, i.e., landmarks, that are both human readable and labeled. Its integration with an explanation methodology, the Layer-wise Relevance Propagation, supports here the automatic compilation of argumentations for the Kernel-based Deep Architecture decisions, expressed in form of analogy with activated landmarks. Quantitative evaluation against the Semantic Role Labeling task, both in English and Italian, suggests that explanations based on semantic and syntagmatic structures are rich and characterize convincing arguments, as they effectively help the user in assessing whether or not to trust the machine decisions. |
| format | Article |
| id | doaj-art-22dc315ecfec4dad9057bb01e6ef603d |
| institution | OA Journals |
| issn | 2499-4553 |
| language | English |
| publishDate | 2019-06-01 |
| publisher | Accademia University Press |
| record_format | Article |
| series | IJCoL |
| spelling | doaj-art-22dc315ecfec4dad9057bb01e6ef603d2025-08-20T01:55:06ZengAccademia University PressIJCoL2499-45532019-06-0151113110.4000/ijcol.453On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple LanguagesDaniele RossiniDanilo CroceRoberto BasiliSentence embeddings are effective input vectors for the neural learning of a number of inferences about content and meaning. Unfortunately, most of such decision processes are epistemologically opaque as for the limited interpretability of the acquired neural models based on the involved embeddings. In this paper, we concentrate on the readability of neural models, discussing an embedding technique (the Nyström methodology) that corresponds to the reconstruction of a sentence in a kernel space, capturing grammatical and lexical semantic information. From this method, we build a Kernel-based Deep Architecture that is characterized by inherently high interpretability properties, as the proposed embedding is derived from examples, i.e., landmarks, that are both human readable and labeled. Its integration with an explanation methodology, the Layer-wise Relevance Propagation, supports here the automatic compilation of argumentations for the Kernel-based Deep Architecture decisions, expressed in form of analogy with activated landmarks. Quantitative evaluation against the Semantic Role Labeling task, both in English and Italian, suggests that explanations based on semantic and syntagmatic structures are rich and characterize convincing arguments, as they effectively help the user in assessing whether or not to trust the machine decisions.https://journals.openedition.org/ijcol/453 |
| spellingShingle | Daniele Rossini Danilo Croce Roberto Basili On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages IJCoL |
| title | On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages |
| title_full | On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages |
| title_fullStr | On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages |
| title_full_unstemmed | On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages |
| title_short | On the Readability of Kernel-based Deep Learning Models in Semantic Role Labeling Tasks over Multiple Languages |
| title_sort | on the readability of kernel based deep learning models in semantic role labeling tasks over multiple languages |
| url | https://journals.openedition.org/ijcol/453 |
| work_keys_str_mv | AT danielerossini onthereadabilityofkernelbaseddeeplearningmodelsinsemanticrolelabelingtasksovermultiplelanguages AT danilocroce onthereadabilityofkernelbaseddeeplearningmodelsinsemanticrolelabelingtasksovermultiplelanguages AT robertobasili onthereadabilityofkernelbaseddeeplearningmodelsinsemanticrolelabelingtasksovermultiplelanguages |