Constructing explainable classifiers from the start: enabling human-in-the loop machine learning
Interactive machine learning (IML) enables the incorporation of human expertise because the human participates in the construction of the learned model. Moreover, with human-in-the-loop machine learning (HITL-ML), the human experts drive the learning, and they can steer the learning objective not on...
| Authors: | ; ; |
|---|---|
| Format: | Article |
| Status: | Published version |
| Publication date: | 2022 |
| Country: | España |
| Institution: | Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya) |
| Repository: | Recercat. Dipósit de la Recerca de Catalunya |
| OAI Identifier: | oai:recercat.cat:10230/55636 |
| Online Access: | http://hdl.handle.net/10230/55636 http://dx.doi.org/10.3390/info13100464 |
| Access Level: | Open access |
| Keyword: | Interactive machine learning Decision tree classifiers Transparent-by-design Parallel coordinates |
| _version_ | 1878740691395280896 |
|---|---|
| author | Estivill-Castro, V. (Vladimir) |
| author2 | Gilmore, Eugene Hexel, René |
| author2_role | author author |
| author_browse | Estivill-Castro, V. (Vladimir) Gilmore, Eugene Hexel, René |
| author_facet | Estivill-Castro, V. (Vladimir) Gilmore, Eugene Hexel, René |
| author_role | author |
| collection | Recercat. Dipósit de la Recerca de Catalunya |
| description | Interactive machine learning (IML) enables the incorporation of human expertise because the human participates in the construction of the learned model. Moreover, with human-in-the-loop machine learning (HITL-ML), the human experts drive the learning, and they can steer the learning objective not only for accuracy but perhaps for characterisation and discrimination rules, where separating one class from others is the primary objective. Moreover, this interaction enables humans to explore and gain insights into the dataset as well as validate the learned models. Validation requires transparency and interpretable classifiers. The huge relevance of understandable classification has been recently emphasised for many applications under the banner of explainable artificial intelligence (XAI). We use parallel coordinates to deploy an IML system that enables the visualisation of decision tree classifiers but also the generation of interpretable splits beyond parallel axis splits. Moreover, we show that characterisation and discrimination rules are also well communicated using parallel coordinates. In particular, we report results from the largest usability study of a IML system, confirming the merits of our approach. |
| eu_rights_str_mv | openAccess |
| format | article |
| id | ES_d7a63945f611fbdfdfffe623206787c8 |
| institution | Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya) |
| instname_str | Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya) |
| network_acronym_str | ES |
| network_name_str | España |
| oai_identifier_str | oai:recercat.cat:10230/55636 |
| publishDate | 2022 |
| publishDateSort | 2022 |
| publisher | MDPI |
| publisherStr | MDPI |
| reponame_str | Recercat. Dipósit de la Recerca de Catalunya |
| spelling | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning Estivill-Castro, V. (Vladimir) Gilmore, Eugene Hexel, René Interactive machine learning Decision tree classifiers Transparent-by-design Parallel coordinates Interactive machine learning (IML) enables the incorporation of human expertise because the human participates in the construction of the learned model. Moreover, with human-in-the-loop machine learning (HITL-ML), the human experts drive the learning, and they can steer the learning objective not only for accuracy but perhaps for characterisation and discrimination rules, where separating one class from others is the primary objective. Moreover, this interaction enables humans to explore and gain insights into the dataset as well as validate the learned models. Validation requires transparency and interpretable classifiers. The huge relevance of understandable classification has been recently emphasised for many applications under the banner of explainable artificial intelligence (XAI). We use parallel coordinates to deploy an IML system that enables the visualisation of decision tree classifiers but also the generation of interpretable splits beyond parallel axis splits. Moreover, we show that characterisation and discrimination rules are also well communicated using parallel coordinates. In particular, we report results from the largest usability study of a IML system, confirming the merits of our approach. MDPI http://hdl.handle.net/10230/55636 http://dx.doi.org/10.3390/info13100464 |
| spellingShingle | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning Estivill-Castro, V. (Vladimir) Interactive machine learning Decision tree classifiers Transparent-by-design Parallel coordinates |
| status_str | publishedVersion |
| title | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning |
| title_full | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning |
| title_fullStr | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning |
| title_full_unstemmed | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning |
| title_short | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning |
| title_sort | Constructing explainable classifiers from the start: enabling human-in-the loop machine learning |
| topic | Interactive machine learning Decision tree classifiers Transparent-by-design Parallel coordinates |
| topic_facet | Interactive machine learning Decision tree classifiers Transparent-by-design Parallel coordinates |
| url | http://hdl.handle.net/10230/55636 http://dx.doi.org/10.3390/info13100464 |