{"id":2693,"date":"2022-07-29T08:12:00","date_gmt":"2022-07-29T06:12:00","guid":{"rendered":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/?p=2693"},"modified":"2022-09-27T13:31:04","modified_gmt":"2022-09-27T11:31:04","slug":"automatisches-lernen","status":"publish","type":"post","link":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/","title":{"rendered":"Podcast: Automatisches Lernen"},"content":{"rendered":"

Neue Podcastfolge: Wie k\u00f6nnen Machine Learning Prozesse m\u00f6glichst effizient automatisiert werden?<\/strong><\/p>\n

<\/p>\n

In der neuen Podcastfolge \u00ad\u00ad\u00adspreche ich mit Kompetenzs\u00e4ulenkoordinator Florian Karl \u00fcber die aktuellen Forschungsmethoden, Herausforderungen und Anwendungsm\u00f6glichkeiten der Kompetenzs\u00e4ule \u00bbAutomatisches Lernen\u00ab (AutoML)<\/a> im Ada Lovelace Center for Analytics, Data and Applications.<\/a><\/p>\n

Maschinelles Lernen ist mittlerweile vielen ein Begriff. \u00a0Dabei geht es nicht darum, Computer zu programmieren, sondern diese aus vorhandenen Daten selbstst\u00e4ndig lernen zu lassen. Dazu bedarf es bestimmter Modelle, die f\u00fcr die zur Verf\u00fcgung stehenden Daten geeignet sind. Die Modellauswahl, die Konfiguration und auch das Testen der Modelle \u00fcbernehmen bisher Machine Learning-Experten und Data Scientists. Dies ist mit viel manuellem Aufwand verbunden, da z.B. jedes Machine Learning-Modell \u00fcber eine gewisse Anzahl an Hyperparamatern verf\u00fcgt, die alle Einfluss auf die Performance haben k\u00f6nnen. Genau hier versucht AutoML Abhilfe zu schaffen: AutoML ist eine Meta-Methode, die auf sehr viele und unterschiedliche Problemstellungen und Datensituationen angewendet werden kann, um die beschriebenen Machine Learning-Prozessschritte zu automatisieren und damit Wissenschaftler und Data Scientists zu entlasten sowie Machine Learning-Methoden zug\u00e4nglicher zu machen.<\/p>\n

AutoML Systeme m\u00fcssen aber nat\u00fcrlich so konzipiert werden, dass sie bestimmte Arten von Daten unterst\u00fctzen k\u00f6nnen. Unterschiedliche Datentypen erfordern verschiedenste Operationen (z.B. zur Verarbeitung der Daten) und oft auch unterschiedliche Modelle. Gerade spezielle Datentypen wie z.B. Zeitreihendaten, die wir im Interview mit Christopher Mutschler zum sequenzbasierten Lernen<\/a> erw\u00e4hnt haben, oder multimodale Daten (also z.B. Text und Bild in Kombination) stellen bestehende AutoML-L\u00f6sungen vor gro\u00dfe Herausforderungen, da sehr spezielle Methoden erforderlich sind.<\/p>\n

Generell ist dies eine grunds\u00e4tzliche Herausforderung bei der Erstellung eines AutoML-Systems: Zu definieren, was das AutoML-System \u201eausprobieren\u201c darf, um eine gute Konfiguration zu finden. Im Prinzip gilt hier: gibt man dem System nur wenige M\u00f6glichkeiten zum Ausprobieren, geschieht die Suche effizient und findet schnell eine optimale L\u00f6sung, m\u00f6glicherweise werden aber einige gute Optionen weggelassen. Gibt man dem AutoML-System die M\u00f6glichkeit zahlreiche verschiedenen Modelle, Operationen etc. auszuprobieren, ist die Suche langwierig und vielleicht auch nicht so robust.<\/p>\n

Wenn Sie also wissen wollen, wie Prozesse effizient automatisiert werden k\u00f6nnen, ohne dass Experten manuell Modelle ausw\u00e4hlen, konfigurieren und testen m\u00fcssen, h\u00f6ren Sie unbedingt rein!<\/p>\nhttps:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-content\/uploads\/2022\/07\/ADAwillswissen_Podcast_Automatisches.mp3<\/a><\/audio>\n

\n

Automated Adaptive Learning<\/strong><\/p>\n

How can machine learning processes be automated as efficiently as possible?<\/strong><\/p>\n

In this new podcast episode I talk to competence pillar coordinator Florian Karl about the current research methods, challenges and application possibilities of the competence pillar \u00bbAutomated Adaptive Learning\u00ab (AutoML) at the Ada Lovelace Center for Analytics, Data and Applications.<\/p>\n

Everyone is now familiar with machine learning.\u00a0 AutoML is not about programming computers, but letting them learn on their own from existing data. This requires certain models that are suitable for the existing data. Until now, machine learning experts and data scientists have been responsible for selecting, configuring and testing the models. This involves a lot of manual effort because, for example, each machine learning model has a certain number of hyperparameters, which can affect the performance. This is exactly where AutoML comes into play: AutoML is a meta-method that can be applied to a very large number and variety of problems and data situations in order to automate the machine learning process steps. The goal is to reduce the workload of experts and data scientists and to make machine learning methods more accessible.<\/p>\n

AutoML systems, however, must of course be designed to support specific data types. Different data types require different operations (e.g., to process the data) and often different models. In particular, special data types such as time series data, which we mentioned in our interview with Christopher Mutschler about sequence-based learning, or multimodal data (i.e., text and image in combination) pose major challenges for existing AutoML solutions because they require very specific methods.<\/p>\n

In general, this poses a fundamental challenge to building an AutoML system: defining what the AutoML system is allowed to “try” in order to find a good configuration. If one gives the system only a few options to try, the search is in principle efficient and quickly finds an optimal solution, but some good options may have been be left out. If one allows the AutoML system to try numerous different models, operations, etc., the search will be lengthy and perhaps not as robust.<\/p>\n

If you want to know how processes can be automated efficiently without experts having to manually select, configure and test models, listen the new episode.<\/p>\n

https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-content\/uploads\/2022\/07\/Podcast_Automatisches_Lernen.mp4<\/a><\/video><\/div>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

Neue Podcastfolge: Wie k\u00f6nnen Machine Learning Prozesse m\u00f6glichst effizient automatisiert werden?<\/p>\n","protected":false},"author":2,"featured_media":2730,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[54,1,55],"tags":[96,95,98,72,84,85],"class_list":["post-2693","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-anwendungen","category-forschung","category-menschen","tag-automated-learning","tag-automatisches-lernen","tag-automl","tag-machine-learning","tag-maschinelles-lernen","tag-ml-verfahren"],"acf":[],"yoast_head":"\nPodcast: Automatisches Lernen - adalovelacecenter-blog<\/title>\n<meta name=\"description\" content=\"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/\" \/>\n<meta property=\"og:locale\" content=\"de_DE\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"\u00bbADA wills wissen\u00ab Podcast\" \/>\n<meta property=\"og:description\" content=\"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten Florian Karl aus dem ADA Lovelace Center: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?\" \/>\n<meta property=\"og:url\" content=\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/\" \/>\n<meta property=\"og:site_name\" content=\"adalovelacecenter-blog\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/de-de.facebook.com\/FraunhoferIIS\" \/>\n<meta property=\"article:published_time\" content=\"2022-07-29T06:12:00+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2022-09-27T11:31:04+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-content\/uploads\/2022\/06\/ADA_Loveladecenter_Automatisches_Lernen-1.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1367\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Anik\u00f3 Enderlein\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:title\" content=\"\u00bbADA wills wissen\u00ab Podcast\" \/>\n<meta name=\"twitter:description\" content=\"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten Florian Karl aus dem ADA Lovelace Center: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-content\/uploads\/2022\/06\/ADA_Loveladecenter_Automatisches_Lernen-1.png\" \/>\n<meta name=\"twitter:creator\" content=\"@FraunhoferIIS\" \/>\n<meta name=\"twitter:site\" content=\"@FraunhoferIIS\" \/>\n<meta name=\"twitter:label1\" content=\"Verfasst von\" \/>\n\t<meta name=\"twitter:data1\" content=\"Anik\u00f3 Enderlein\" \/>\n\t<meta name=\"twitter:label2\" content=\"Gesch\u00e4tzte Lesezeit\" \/>\n\t<meta name=\"twitter:data2\" content=\"4\u00a0Minuten\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/\",\"url\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/\",\"name\":\"Podcast: Automatisches Lernen - adalovelacecenter-blog\",\"isPartOf\":{\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#website\"},\"datePublished\":\"2022-07-29T06:12:00+00:00\",\"dateModified\":\"2022-09-27T11:31:04+00:00\",\"author\":{\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#\/schema\/person\/db1f6550f5aeffe2718604493fee7639\"},\"description\":\"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?\",\"breadcrumb\":{\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/#breadcrumb\"},\"inLanguage\":\"de\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Startseite\",\"item\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Podcast: Automatisches Lernen\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#website\",\"url\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/\",\"name\":\"adalovelacecenter-blog\",\"description\":\"Kompetenzzentrum f\u00fcr Analytics, Daten und Applikationen\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"de\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#\/schema\/person\/db1f6550f5aeffe2718604493fee7639\",\"name\":\"Anik\u00f3 Enderlein\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/942081197915d7b4a2d8f7bfc7b23d11dd4ecb2dfd82459877ad80d63ab5cc9e?s=96&d=blank&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/942081197915d7b4a2d8f7bfc7b23d11dd4ecb2dfd82459877ad80d63ab5cc9e?s=96&d=blank&r=g\",\"caption\":\"Anik\u00f3 Enderlein\"},\"url\":\"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/author\/aniko_enderlein\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Podcast: Automatisches Lernen - adalovelacecenter-blog","description":"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/","og_locale":"de_DE","og_type":"article","og_title":"\u00bbADA wills wissen\u00ab Podcast","og_description":"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten Florian Karl aus dem ADA Lovelace Center: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?","og_url":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/","og_site_name":"adalovelacecenter-blog","article_publisher":"https:\/\/de-de.facebook.com\/FraunhoferIIS","article_published_time":"2022-07-29T06:12:00+00:00","article_modified_time":"2022-09-27T11:31:04+00:00","og_image":[{"width":1367,"height":768,"url":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-content\/uploads\/2022\/06\/ADA_Loveladecenter_Automatisches_Lernen-1.png","type":"image\/png"}],"author":"Anik\u00f3 Enderlein","twitter_card":"summary_large_image","twitter_title":"\u00bbADA wills wissen\u00ab Podcast","twitter_description":"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten Florian Karl aus dem ADA Lovelace Center: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?","twitter_image":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-content\/uploads\/2022\/06\/ADA_Loveladecenter_Automatisches_Lernen-1.png","twitter_creator":"@FraunhoferIIS","twitter_site":"@FraunhoferIIS","twitter_misc":{"Verfasst von":"Anik\u00f3 Enderlein","Gesch\u00e4tzte Lesezeit":"4\u00a0Minuten"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/","url":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/","name":"Podcast: Automatisches Lernen - adalovelacecenter-blog","isPartOf":{"@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#website"},"datePublished":"2022-07-29T06:12:00+00:00","dateModified":"2022-09-27T11:31:04+00:00","author":{"@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#\/schema\/person\/db1f6550f5aeffe2718604493fee7639"},"description":"Neue Podcastfolge zum Automatischen Lernen mit unserem Experten: Wie lernen Modelle aus vorhandenen Daten selbstst\u00e4ndig?","breadcrumb":{"@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/#breadcrumb"},"inLanguage":"de","potentialAction":[{"@type":"ReadAction","target":["https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/automatisches-lernen\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Startseite","item":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/"},{"@type":"ListItem","position":2,"name":"Podcast: Automatisches Lernen"}]},{"@type":"WebSite","@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#website","url":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/","name":"adalovelacecenter-blog","description":"Kompetenzzentrum f\u00fcr Analytics, Daten und Applikationen","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"de"},{"@type":"Person","@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#\/schema\/person\/db1f6550f5aeffe2718604493fee7639","name":"Anik\u00f3 Enderlein","image":{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/942081197915d7b4a2d8f7bfc7b23d11dd4ecb2dfd82459877ad80d63ab5cc9e?s=96&d=blank&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/942081197915d7b4a2d8f7bfc7b23d11dd4ecb2dfd82459877ad80d63ab5cc9e?s=96&d=blank&r=g","caption":"Anik\u00f3 Enderlein"},"url":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/author\/aniko_enderlein\/"}]}},"_links":{"self":[{"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/posts\/2693","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/comments?post=2693"}],"version-history":[{"count":10,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/posts\/2693\/revisions"}],"predecessor-version":[{"id":2764,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/posts\/2693\/revisions\/2764"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/media\/2730"}],"wp:attachment":[{"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/media?parent=2693"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/categories?post=2693"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/websites.fraunhofer.de\/adalovelacecenter-blog\/wp-json\/wp\/v2\/tags?post=2693"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}