eXplainable Deep Neural Networks (xDNN)<\/span><\/h3>\n![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem1-300x145.png\")
<\/p>\n
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem3-300x150.png\")
<\/p>\n
<\/p>\n
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem2-300x198.png\")
<\/td>\n
<\/p>\nxDNN<\/a> is a prototype-based network that uses data density as its core mechanism.<\/span><\/p>\nPrototypes are selected data samples that the users can easily view, understand and analyze their similarity to other data samples.<\/span><\/p>\nRelated papers:<\/span>\u00a0<\/span><\/strong><\/p>\nAngelov, P., & Soares, E. (2020). Towards explainable deep neural networks (xDNN). Neural Networks, 130, 185-194.<\/span><\/a><\/span><\/p>\nSoares, E., Angelov, P., Costa, B., & Castro, M. (2019, July). Actively semi-supervised deep rule-based classifier applied to adverse driving scenarios. In\u00a02019 <\/i>International<\/i> Joint <\/i>Conference<\/i> on<\/i> Neural Networks (IJCNN)<\/i>\u00a0(pp. 1-8). IEEE.<\/a><\/p>\nGit:<\/span><\/p>\nhttps:\/\/github.com\/Plamen-Eduardo\/xDNN—Python<\/a><\/span><\/p>\n<\/td>\n<\/tr>\n\n\n<\/h3>\nxDNN<\/b> for model interpretability in the context of DRL<\/b><\/span><\/h3>\n <\/p>\n
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem4-300x126.png\")
<\/p>\n
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem5-300x130.png\")
<\/td>\n
\n\nxDNN<\/a> is used to approximate the DRL model with a set of IF\u2026THEN rules that provide an alternative interpretable model, which is further enhanced by visualizing the rules.<\/span>\u200b<\/span><\/span><\/span><\/p>\nRelated papers:<\/span>\u00a0<\/span><\/strong><\/p>\n\u200b<\/span><\/span>Soares, E. A., Angelov, P. P., Costa, B., Castro, M., Nageshrao, S., & Filev, D. (2020). Explaining deep learning models through rule-based approximation and visualization. IEEE Transactions on Fuzzy Systems.<\/a><\/span><\/p>\nSoares, E., Angelov, P., Filev, D., Costa, B., Castro, M., & Nageshrao, S. (2019, December). Explainable density-based approach for self-driving actions classification. In 2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA) (pp. 469-474). IEEE.<\/a><\/span><\/p>\n<\/div>\n<\/div>\n<\/td>\n<\/tr>\n\n\nxDNN<\/b> for object detection<\/b><\/span><\/h3>\n![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem6-300x85.png\")
<\/p>\n
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Captura-de-Tela-2021-09-24-a\u0300s-09.23.43-300x203.png\")
<\/td>\n
xDNN<\/a> is applied for the object detection context. Due to its transparent design, it allows an explainable error analysis which is crucial for high stake applications.\u00a0 <\/span><\/td>\n<\/tr>\n\n\nxClass<\/b><\/span> for Novelty detection<\/span> <\/b><\/h3>\n <\/p>\n
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/08\/detection_noise-300x169.gif\")
![\"\"](\"http:\/\/wp.lancs.ac.uk\/xai\/files\/2021\/09\/Imagem8-300x164.png\")
<\/td>\n
xClass<\/a> extends xDNN and offers:<\/span><\/p>\n\n- Real-time \u201cnovelty detection\u201d through the recursive density and the m<\/i>-sigma rule.<\/span><\/li>\n
- Weakly supervised structure that allows few 1 class and few data samples initially.<\/span><\/li>\n
- Automatic creation of new classes.<\/span><\/li>\n
- Transparent structure that allows humans to audit it.<\/span><\/li>\n<\/ul>\n
Related papers:<\/span>\u00a0<\/span><\/strong><\/p>\nAngelov, P., & Soares, E. (2021). Detecting and learning from unknown by extremely weak supervision: exploratory classifier (xClass).\u00a0Neural Computing and Applications<\/i>, 1-13.<\/a><\/p>\nSoares, E., Angelov, P., Costa, B., & Castro, M. (2019, July). Actively semi-supervised deep rule-based classifier applied to adverse driving scenarios. In\u00a02019 <\/i>International<\/i> Joint <\/i>Conference<\/i> on<\/i> Neural Networks (IJCNN)<\/i>\u00a0(pp. 1-8). IEEE.<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n","protected":false},"excerpt":{"rendered":" eXplainable Deep Neural Networks (xDNN) xDNN is a prototype-based network that uses data density as its core mechanism. Prototypes are selected data samples that the users can easily view, understand and analyze their similarity to other data samples. Related papers:\u00a0 Angelov, P., & Soares, E. (2020). Towards explainable deep neural networks (xDNN).… Read More »Applications<\/span><\/a><\/p>\n","protected":false},"author":1447,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"off","neve_meta_content_width":100,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","footnotes":""},"class_list":["post-63","page","type-page","status-publish","hentry"],"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages\/63","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/users\/1447"}],"replies":[{"embeddable":true,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/comments?post=63"}],"version-history":[{"count":14,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages\/63\/revisions"}],"predecessor-version":[{"id":88,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages\/63\/revisions\/88"}],"wp:attachment":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/media?parent=63"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
<\/p>\n<\/p>\n<\/td>\nxDNN<\/a> is a prototype-based network that uses data density as its core mechanism.<\/span><\/p>\n Prototypes are selected data samples that the users can easily view, understand and analyze their similarity to other data samples.<\/span><\/p>\n Related papers:<\/span>\u00a0<\/span><\/strong><\/p>\n Angelov, P., & Soares, E. (2020). Towards explainable deep neural networks (xDNN). Neural Networks, 130, 185-194.<\/span><\/a><\/span><\/p>\n Soares, E., Angelov, P., Costa, B., & Castro, M. (2019, July). Actively semi-supervised deep rule-based classifier applied to adverse driving scenarios. In\u00a02019 <\/i>International<\/i> Joint <\/i>Conference<\/i> on<\/i> Neural Networks (IJCNN)<\/i>\u00a0(pp. 1-8). IEEE.<\/a><\/p>\n Git:<\/span><\/p>\n https:\/\/github.com\/Plamen-Eduardo\/xDNN—Python<\/a><\/span><\/p>\n<\/td>\n<\/tr>\n <\/p>\n xDNN<\/a> is used to approximate the DRL model with a set of IF\u2026THEN rules that provide an alternative interpretable model, which is further enhanced by visualizing the rules.<\/span>\u200b<\/span><\/span><\/span><\/p>\n Related papers:<\/span>\u00a0<\/span><\/strong><\/p>\n \u200b<\/span><\/span>Soares, E. A., Angelov, P. P., Costa, B., Castro, M., Nageshrao, S., & Filev, D. (2020). Explaining deep learning models through rule-based approximation and visualization. IEEE Transactions on Fuzzy Systems.<\/a><\/span><\/p>\n Soares, E., Angelov, P., Filev, D., Costa, B., Castro, M., & Nageshrao, S. (2019, December). Explainable density-based approach for self-driving actions classification. In 2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA) (pp. 469-474). IEEE.<\/a><\/span><\/p>\n<\/div>\n <\/p>\n Related papers:<\/span>\u00a0<\/span><\/strong><\/p>\n Angelov, P., & Soares, E. (2021). Detecting and learning from unknown by extremely weak supervision: exploratory classifier (xClass).\u00a0Neural Computing and Applications<\/i>, 1-13.<\/a><\/p>\n Soares, E., Angelov, P., Costa, B., & Castro, M. (2019, July). Actively semi-supervised deep rule-based classifier applied to adverse driving scenarios. In\u00a02019 <\/i>International<\/i> Joint <\/i>Conference<\/i> on<\/i> Neural Networks (IJCNN)<\/i>\u00a0(pp. 1-8). IEEE.<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n","protected":false},"excerpt":{"rendered":" eXplainable Deep Neural Networks (xDNN) xDNN is a prototype-based network that uses data density as its core mechanism. Prototypes are selected data samples that the users can easily view, understand and analyze their similarity to other data samples. Related papers:\u00a0 Angelov, P., & Soares, E. (2020). Towards explainable deep neural networks (xDNN).… Read More »Applications<\/span><\/a><\/p>\n","protected":false},"author":1447,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"off","neve_meta_content_width":100,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","footnotes":""},"class_list":["post-63","page","type-page","status-publish","hentry"],"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages\/63","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/users\/1447"}],"replies":[{"embeddable":true,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/comments?post=63"}],"version-history":[{"count":14,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages\/63\/revisions"}],"predecessor-version":[{"id":88,"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/pages\/63\/revisions\/88"}],"wp:attachment":[{"href":"https:\/\/wp.lancs.ac.uk\/xai\/wp-json\/wp\/v2\/media?parent=63"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\n \n <\/h3>\n
xDNN<\/b> for model interpretability in the context of DRL<\/b><\/span><\/h3>\n
<\/p>\n<\/td>\n\n \n \n xDNN<\/b> for object detection<\/b><\/span><\/h3>\n
<\/p>\n<\/td>\nxDNN<\/a> is applied for the object detection context. Due to its transparent design, it allows an explainable error analysis which is crucial for high stake applications.\u00a0 <\/span><\/td>\n<\/tr>\n \n \n xClass<\/b><\/span> for Novelty detection<\/span> <\/b><\/h3>\n
<\/td>\nxClass<\/a> extends xDNN and offers:<\/span><\/p>\n \n