{"id":25020,"date":"2019-08-08T00:00:00","date_gmt":"2024-09-13T08:48:09","guid":{"rendered":"https:\/\/wordpress-583806-4798031.cloudwaysapps.com\/twin-examples-of-multiple-trees-1-uml-models-2-machine-learning\/"},"modified":"2024-11-15T16:21:53","modified_gmt":"2024-11-15T15:21:53","slug":"twin-examples-of-multiple-trees-1-uml-models-2-machine-learning","status":"publish","type":"post","link":"https:\/\/informator.se\/en\/twin-examples-of-multiple-trees-1-uml-models-2-machine-learning\/","title":{"rendered":"Twin examples of multiple trees: 1. UML models, 2. Machine Learning"},"content":{"rendered":"\n
\n
\n
\n
\n
\n
\n
\n
\u201cToday, professionals get trained in using tools\u2026 \u00a0there\u2019s a lack of education of fundamentals like modeling, architecture, methods, or concepts… Getting value out of data needs professionalization based on education and practical experience.\u201d\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0<\/span><\/i><\/div>\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0<\/span><\/i><\/div>\n
Andreas Buckenhofer, Daimler <\/span>TSS<\/a><\/b>, in an interview by <\/span>OODBMS<\/a> on Big Data, July 2019\u00a0<\/span>
<\/span><\/div>\n
\u00a0<\/div>\n
In my opinion, he\u2019s spot on.<\/span><\/div>\n
\u00a0<\/span><\/div>\n
My post<\/span><\/a> from March mentions why new AI languages aren\u2019t exactly heavies of a CV in a mainstream business; in April, a figure (at the end of the post) also touched on Forest structures in ML and eXplainable AI.<\/span><\/div>\n
\u00a0<\/span><\/div>\n
After a free-wind sail that took us from detail to architecture, we now go into some structural \u201cforestry\u201d. It\u2019s about tackling the same domain from multiple viewpoints, instead of clinging on to one.<\/span><\/div>\n

1. Multiple trees in UML: Generalization sets<\/span><\/b><\/h3>\n
My post<\/span> from 2015 (in Swedish) discusses the sets in more detail, so let\u2019s just recap the diagrams, in English, and add \u00abpowertype\u00bb<\/b> on the fourth one (a<\/span><\/span> power set\u2019s instances are subsets, so by the same token, a UML2 powertype\u2019s instances are \u201csubtypes\u201d of a general construct).<\/span><\/div>\n
<\/a><\/div>\n
\n
<\/a><\/div>\n
<\/a><\/div>\n

\u00a0<\/p>\n

<\/a><\/div>\n<\/div>\n

2. Multiple trees in Machine Learning: random Decision Forests<\/span><\/b><\/h3>\n
Here, a designer is not the one who maps out subclasses. Rather, an ML algorithm generates in training time, from (labeled) data, an ability to perform classification (i.e accurate \u201cmapping-out\u201d of \u201cclasses\u201d). The decision nodes of a (classification) tree gradually subdivide the data into more and more fine-grained classes.<\/span><\/div>\n
\u00a0<\/div>\n
Why bother about decision trees when Deep Neural Networks are booming? Because explainability<\/span><\/a> opens the door to acceptance in mission-critical apps. ML-generated logic has to be auditable. User enterprises are pushing for graphicness, conceptualization, traceability, V&V. Those are the strengths of decision trees, and weaknesses of Deep NNs (we know those work, but hardly how); same thing with learning time required, size of training data sets, execution speed, or partitionability (a hint for IT architects: a tree works independently, whereas a neuron relies on many other ones). Atop of that, decision trees offer a structural backbone of hybrid AI systems (see also the last paragraphs in this post<\/span><\/a>).<\/span><\/div>\n
\u00a0<\/div>\n
You might remember that trees in woods<\/span><\/a> sometimes fuse their roots and exchange materials. Unsurprisingly, we find some synergy in virtual forests too. Firstly, our <\/span>trees are \u201cgrown\u201d on a random sample<\/b> each (hence some \u201cbiodiversity\u201d too), from one training-data set (hence fewer terabytes of training data). Secondly, on each sample, its tree\u2019s decision nodes use a random<\/b> subset<\/b> of all available attributes. This gives architects and other roles some room to tune the mix of efficiency and explainability; in forests, it\u2019s is near the level of genetic algorithms (GAs too have possible \u201cmix-tuning points\u201d in \u201cbiodiversity steps\u201d Crossover and Mutation).<\/span><\/a><\/div>\n
\u00a0<\/div>\n
The more trees and \u201cbiodiversity\u201d our ML algorithm grows, the more accurate and robust the generated logic becomes, because the final step is vote counting. A forest\u2019s output (a classification like here, or a forecast) is an aggregated value of the outputs of all trees (a statistical mode in classification, or a mean in regression). It prevents the random decision forest from getting stuck in local optima, that is, we minimize error rates and overfitting to a given training-data set (which may be both incomplete and biased).<\/span><\/div>\n<\/div><\/div>\n<\/div><\/div>\n<\/div><\/div>\n<\/div><\/div>\n<\/div><\/div>\n<\/div><\/div>\n<\/div><\/div>\n<\/span><\/span>\r\n

Milan Kratochvil\u00a0<\/span><\/span><\/h4>\r\nTrainer at Informator, senior modeling and architecture consultant at Kiseldalens, main author : UML Extra Light (Cambridge University Press) and Growing Modular (Springer), Advanced UML2 Professional (OCUP<\/a> cert level 3\/3).<\/em>\n\nMilan and Informator collaborate since 1996 on architecture, modelling, UML, requirements, rules, and design.\u00a0<\/em><\/span><\/span>","protected":false},"excerpt":{"rendered":"

\u201cToday, professionals get trained in using tools\u2026 \u00a0there\u2019s a lack of education of fundamentals like modeling, architecture, methods, or concepts… Getting value out of data needs professionalization based on education and practical experience.\u201d\u00a0\u00a0\u00a0<\/p>\n","protected":false},"author":6,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"footnotes":""},"categories":[571,570,922],"tags":[591,592],"class_list":["post-25020","post","type-post","status-publish","format-standard","hentry","category-ai","category-blogg","category-blogi","tag-ai","tag-machine-learning"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t