Webin\u00e1\u0159 s Dominikem V\u00edtem, lektorem spole\u010dnosti Datascript<\/a>,\u00a0 po\u0159\u00e1d\u00e1me v r\u00e1mci vzd\u011bl\u00e1vac\u00edho festivalu MIXER.<\/strong><\/p>\n

PRO\u010c SE P\u0158IHL\u00c1SIT?<\/strong>
\nHlubok\u00e9 neuronov\u00e9 s\u00edt\u011b dnes ve v\u011bt\u0161in\u011b \u00faloh spojen\u00fdch s Computer vision (zpracov\u00e1n\u00ed obrazu a strojov\u00e9 vid\u011bn\u00ed) jednozna\u010dn\u011b dominuj\u00ed nad klasick\u00fdmi p\u0159\u00edstupy i nad lidsk\u00fdmi rozli\u0161ovac\u00edmi schopnostmi. \u010casto je v\u0161ak prvotn\u00ed nad\u0161en\u00ed z klasifik\u00e1toru ps\u016f a ko\u010dek (nejtypi\u010dt\u011bj\u0161\u00ed demo p\u0159\u00edklad) vyst\u0159\u00edd\u00e1no strachem ze skute\u010dn\u00e9ho nasazen\u00ed do produkce nebo neporozum\u011bn\u00ed, jak\u00fd p\u0159\u00edstup, model nebo architekturu pou\u017e\u00edt na konkr\u00e9tn\u00ed probl\u00e9m \u010di zad\u00e1n\u00ed.<\/p>\n

AI, Machine Learning, Deep Learning – sezn\u00e1men\u00ed s\u202fpojmy a v\u00fdvojem discipl\u00edn, rozd\u00edly mezi nimi, pou\u017e\u00edvan\u00e1 terminologie.
\nZku\u0161enosti s \u00fasp\u011b\u0161n\u00fdmi i ne\u00fasp\u011b\u0161n\u00fdmi projekty zalo\u017een\u00fdmi na deep learningu.
\nNej\u010dast\u011bj\u0161\u00ed probl\u00e9my p\u0159i v\u00fdvoji a nasazen\u00ed \u2013 z\u00edsk\u00e1n\u00ed anotovan\u00e9ho datasetu, tr\u00e9nov\u00e1n\u00ed a inference on-prem nebo v cloudu.
\nP\u0159esnost versus rychlost p\u0159i nasazen\u00ed \u2013 volba spr\u00e1vn\u00e9 architektury pro cloud nebo pro mobiln\u00ed za\u0159\u00edzen\u00ed (Transformers, EfficientNet, YOLO).
\nKomunikace s klientem, businessov\u00e9 metriky, interpretovatelnost.<\/p>\n

Akce je ZDARMA. Sta\u010d\u00ed se pouze t\u011b\u0161it &\u00a0registrovat<\/a>!<\/p>\n

N\u011aCO O SPEAKEROVI
\n<\/strong>Dominik pracuje ve spole\u010dnosti DataSentics a.s. jako machine learning engineer. Zku\u0161enosti s rozpozn\u00e1n\u00edm obrazu a deep learningem z\u00edskal zejm\u00e9na na projektech automatizace ve \u0160koda auto, detekce zbo\u017e\u00ed v oblasti retail nebo automatick\u00e9ho \u010dten\u00ed skenovan\u00fdch dokument\u016f a faktur.<\/p>\n

MIXER<\/strong>\u00a0je vzd\u011bl\u00e1vac\u00ed festival, kter\u00fd s l\u00e1skou a nad\u0161en\u00edm po\u0159\u00e1d\u00e1me nejen pro na\u0161e zam\u011bstnance, ale tak\u00e9 pro v\u0161echny, kte\u0159\u00ed si cht\u011bj\u00ed na ja\u0159e pro\u010distit hlavu, zacvi\u010dit si, dozv\u011bd\u011bt se n\u011bco nov\u00e9ho z r\u016fzn\u00fdch oblast\u00ed IT, zdrav\u00e9ho \u017eivotn\u00edho stylu anebo se t\u0159eba jen pobavit.
\nLetos prob\u00edh\u00e1 od 14.-18. 3. on-line i offline.
\nV\u0161echny on-line p\u0159edn\u00e1\u0161ky jsou p\u0159\u00edstupn\u00e9 ve\u0159ejnosti a najdete je zde na t\u011bchto str\u00e1nk\u00e1ch v sekci MEETUPY<\/a>.<\/p>\n

\u00a0<\/strong><\/p>\n

 <\/p>\n

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

Webin\u00e1\u0159 s Dominikem V\u00edtem, lektorem spole\u010dnosti Datascript,\u00a0 po\u0159\u00e1d\u00e1me v r\u00e1mci vzd\u011bl\u00e1vac\u00edho festivalu MIXER. PRO\u010c SE P\u0158IHL\u00c1SIT? Hlubok\u00e9 neuronov\u00e9 s\u00edt\u011b dnes ve v\u011bt\u0161in\u011b \u00faloh spojen\u00fdch s Computer vision (zpracov\u00e1n\u00ed obrazu a strojov\u00e9 vid\u011bn\u00ed) jednozna\u010dn\u011b dominuj\u00ed nad klasick\u00fdmi p\u0159\u00edstupy i nad lidsk\u00fdmi rozli\u0161ovac\u00edmi schopnostmi. \u010casto je v\u0161ak prvotn\u00ed nad\u0161en\u00ed z klasifik\u00e1toru ps\u016f a ko\u010dek (nejtypi\u010dt\u011bj\u0161\u00ed demo p\u0159\u00edklad) […]<\/p>\n","protected":false},"author":12,"featured_media":242635,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_et_pb_use_builder":"","_et_pb_old_content":"","_et_gb_content_width":"","ghostkit_customizer_options":"","ghostkit_custom_css":"","ghostkit_custom_js_head":"","ghostkit_custom_js_foot":"","ghostkit_typography":"","_vp_format_video_url":"","_vp_image_focal_point":[]},"tags":[],"mec_category":[],"yoast_head":"\n