Feed | mysite

All Posts

Oct 1
1 min

Recipe for Distributed Hyperparameter Search for an unlikely couple: Keras Tuner and Kubeflow

Kubeflow offers a builtin Hyperparameter Search (HS) 'Katib', there are still a number of reasons one may want to prefer Keras Tuner as H...

3 viewsWrite a comment

Kubeflow makes more sense when you understand this. Moving Data Science to Kubeflow in four steps

Roman Kazinnik

Jul 17
3 min

Kubeflow makes more sense when you understand this. Moving Data Science to Kubeflow in four steps

Previously I explained deploying and utilizing Machine Learning Platform concepts and its goals: https://www.romankazinnik.com/post/autom...

96 viewsWrite a comment

Machine Learning as a Flow Cont-ed: Kubeflow vs. Metaflow

Roman Kazinnik

Feb 24
3 min

Machine Learning as a Flow Cont-ed: Kubeflow vs. Metaflow

After having both tools both in our dev and production, here is my comparative analysis of Kubeflow (2018, Google) and Metaflow (2019, Ne...

248 viewsWrite a comment

Distributed Modeling: Ring-Reduce vs. All-Reduce

Roman Kazinnik

Jan 3
2 min

Distributed Modeling: Ring-Reduce vs. All-Reduce

In this blog post, I’d like to share some of the insights from my work at the High-Performance Computing (HPC) Texas Advanced Computing C...

62 viewsWrite a comment

Roman Kazinnik

Aug 23, 2019
1 min

Convolutional Neural Networks for NLP

The goal of this post is to compare NLP approaches for encoding words: (1) bag of words (2) tfidf (3) google GloVe emeddings Training set...

24 viewsWrite a comment

Automate Machine Learning as a Flow: Kubeflow

Roman Kazinnik

Jun 26, 2019
2 min

Automate Machine Learning as a Flow: Kubeflow

I will explain the most recent trends in Machine Learning Automation as a Flow. Examples of ML Flow include Kubeflow by Google, MLFlow, M...

553 viewsWrite a comment

Beyond Supervised ML: learning latent distributions

Roman Kazinnik

Feb 17, 2019
2 min

Beyond Supervised ML: learning latent distributions

Supervised ML is a powerful tool, however there is a range of problems where its utilization can't be justified. Such problems include re...

26 viewsWrite a comment

Convolutional Neural Networks: insights on multi-scale representation and redundant architecture

Roman Kazinnik

Jun 26, 2018
2 min

Convolutional Neural Networks: insights on multi-scale representation and redundant architecture

https://github.com/romanonly/romankazinnik_blog/tree/master/CVision/CNN Perhaps the greatest about CNN (Convolutional Neural Networks) is...

71 viewsWrite a comment

Beyond DAG: Introducing Cyclic Graphs in Neural Networks

Roman Kazinnik

Apr 11, 2018
2 min

Beyond DAG: Introducing Cyclic Graphs in Neural Networks

Deep Learning is driven by the possibility of computing at-scale gradient of the minimization problem for Directed Acyclic Graphs (DAG). ...

189 viewsWrite a comment

Approximation theory, AI, and how they actually go together (talk at UNCC)

Roman Kazinnik

Jan 14, 2018
1 min

Approximation theory, AI, and how they actually go together (talk at UNCC)

Here is a video of my talk at the University of North Carolina at Charlotte (01.12.2018). I show how at the beginning of AI, Deep Learni...

83 viewsWrite a comment

Beyond Supervised ML: which features are really important? Trial reveals surprising patterns

Roman Kazinnik

Jan 12, 2018
7 min

Beyond Supervised ML: which features are really important? Trial reveals surprising patterns

Trial and error reveal surprising patterns of success—and failure—using different BL and Supervised ML models. When we consider ML (Machi...

39 viewsWrite a comment

An Insight into Data Quality: More Data Equals More Work, But Not Necessarily a Better Model

Roman Kazinnik

Jan 8, 2018
4 min

An Insight into Data Quality: More Data Equals More Work, But Not Necessarily a Better Model

In the field of data quality and predictive modeling, which modeling metrics are best at “telling the truth,” so to speak? How did I come...

41 viewsWrite a comment

{"items":["5f7561e970e5a00017e09cb2","5f11deefa9a25e001751b805","5e533556778e3e0017a0be6c","5e0fb8ba580d850017978a28","5d606af563123700160706e3","5d139292c313330016864dc8","5c68dadac4799c001c3bdbc8","5b30363dbb6cb9036373e717","5a642643cda1f60b69f02589","5a5b831fae35e208445b0fad","5a580fe52b5ad3025860864c","5a52f9af18447409e1c2a4c8"],"styles":{"galleryType":"Columns","groupSize":1,"showArrows":true,"cubeImages":true,"cubeType":"min","cubeRatio":2.2857142857142856,"isVertical":true,"gallerySize":940,"collageAmount":0,"collageDensity":0,"groupTypes":"1","oneRow":false,"imageMargin":16,"galleryMargin":0,"floatingImages":0,"chooseBestGroup":true,"smartCrop":false,"hasThumbnails":false,"enableScroll":true,"isGrid":true,"isSlider":false,"isColumns":false,"isSlideshow":false,"cropOnlyFill":false,"fixedColumns":0,"enableInfiniteScroll":true,"isRTL":false,"minItemSize":50,"rotatingGroupTypes":"","rotatingCubeRatio":"","gallerySliderImageRatio":1.7777777777777777,"numberOfImagesPerRow":3,"numberOfImagesPerCol":1,"groupsPerStrip":0,"borderRadius":0,"boxShadow":0,"gridStyle":0,"mobilePanorama":false,"placeGroupsLtr":false,"viewMode":"preview","thumbnailSpacings":4,"galleryThumbnailsAlignment":"bottom","isMasonry":false,"isAutoSlideshow":false,"slideshowLoop":false,"autoSlideshowInterval":4,"useCustomButton":false,"bottomInfoHeight":0,"titlePlacement":"SHOW_BELOW","galleryHorizontalAlign":"center","galleryTextAlign":"center","galleryVerticalAlign":"center","scrollSnap":false,"itemClick":"expand","fullscreen":true,"allowSocial":true,"allowDownload":false,"allowTitle":true,"allowDescription":false,"loveButton":true,"loveCounter":false,"videoPlay":"hover","scrollAnimation":"NO_EFFECT","scrollDirection":0,"overlayAnimation":"FADE_IN","arrowsPosition":0,"arrowsSize":23,"watermarkOpacity":40,"watermarkSize":40,"useWatermark":true,"watermarkDock":{"top":"auto","left":"auto","right":0,"bottom":0,"transform":"translate3d(0,0,0)"},"loadMoreAmount":"all","defaultShowInfoExpand":1,"allowTitleExpand":true,"allowDescriptionExpand":true,"allowLinkExpand":true,"expandInfoPosition":0,"allowFullscreenExpand":true,"fullscreenLoop":false,"galleryAlignExpand":"left","addToCartBorderWidth":1,"addToCartButtonText":"","slideshowInfoSize":200,"playButtonForAutoSlideShow":false,"allowSlideshowCounter":false,"hoveringBehaviour":"NEVER_SHOW","thumbnailSize":120,"magicLayoutSeed":1,"imageHoverAnimation":"NO_EFFECT","calculateTextBoxHeightMode":"MANUAL","calculateTextBoxWidthMode":"PERCENT","textBoxHeight":310,"textBoxWidth":200,"textBoxWidthPercent":50,"textImageSpace":10,"textBoxBorderRadius":0,"textBoxBorderWidth":0,"textsVerticalPadding":-15,"textsHorizontalPadding":-30,"titleDescriptionSpace":6,"customButtonText":"","customButtonBorderWidth":1,"customButtonBorderRadius":0,"loadMoreButtonText":"","loadMoreButtonBorderWidth":1,"loadMoreButtonBorderRadius":0,"imageInfoType":"ATTACHED_BACKGROUND","itemBorderWidth":1,"itemBorderRadius":0,"itemEnableShadow":false,"itemShadowBlur":20,"itemShadowDirection":135,"itemShadowSize":10,"imageLoadingMode":"BLUR","expandAnimation":"NO_EFFECT","imageQuality":90,"usmToggle":false,"usm_a":0,"usm_r":0,"usm_t":0,"videoSound":false,"videoSpeed":"1","videoLoop":true,"gallerySizeType":"px","gallerySizePx":940,"allowContextMenu":true,"itemBorderColor":{"value":"#CCCACA"},"galleryLayout":2,"modifiedGallerySize":true,"selectedLayout":"2|bottom|1|min|true|0|true","layoutsVersion":2,"selectedLayoutV2":2,"isSlideshowFont":true,"externalInfoHeight":310,"externalInfoWidth":0},"container":{"width":940,"galleryWidth":972,"galleryHeight":1440,"scrollBase":0,"height":null}}

Challenge

ML Infra, Math and In Between. By Roman Kazinnik

sign up for my blog and newsletter

Recipe for Distributed Hyperparameter Search for an unlikely couple: Keras Tuner and Kubeflow

Kubeflow makes more sense when you understand this. Moving Data Science to Kubeflow in four steps

Machine Learning as a Flow Cont-ed: Kubeflow vs. Metaflow

Distributed Modeling: Ring-Reduce vs. All-Reduce

Convolutional Neural Networks for NLP

Automate Machine Learning as a Flow: Kubeflow

Beyond Supervised ML: learning latent distributions

Convolutional Neural Networks: insights on multi-scale representation and redundant architecture

Beyond DAG: Introducing Cyclic Graphs in Neural Networks

Approximation theory, AI, and how they actually go together (talk at UNCC)

Beyond Supervised ML: which features are really important? Trial reveals surprising patterns

An Insight into Data Quality: More Data Equals More Work, But Not Necessarily a Better Model