Roman KazinnikOct 11 minRecipe for Distributed Hyperparameter Search for an unlikely couple: Keras Tuner and KubeflowKubeflow offers a builtin Hyperparameter Search (HS) 'Katib', there are still a number of reasons one may want to prefer Keras Tuner as H...

Roman KazinnikJul 173 minKubeflow makes more sense when you understand this. Moving Data Science to Kubeflow in four stepsPreviously I explained deploying and utilizing Machine Learning Platform concepts and its goals: https://www.romankazinnik.com/post/autom...

Roman KazinnikFeb 243 minMachine Learning as a Flow Cont-ed: Kubeflow vs. MetaflowAfter having both tools both in our dev and production, here is my comparative analysis of Kubeflow (2018, Google) and Metaflow (2019, Ne...

Roman KazinnikJan 32 minDistributed Modeling: Ring-Reduce vs. All-ReduceIn 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...

Roman KazinnikAug 23, 20191 minConvolutional Neural Networks for NLPThe goal of this post is to compare NLP approaches for encoding words: (1) bag of words (2) tfidf (3) google GloVe emeddings Training set...

Roman KazinnikJun 26, 20192 minAutomate 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...

Roman KazinnikFeb 17, 20192 minBeyond 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...

Roman KazinnikJun 26, 20182 minConvolutional Neural Networks: insights on multi-scale representation and redundant architecturehttps://github.com/romanonly/romankazinnik_blog/tree/master/CVision/CNN Perhaps the greatest about CNN (Convolutional Neural Networks) is...

Roman KazinnikApr 11, 20182 minBeyond DAG: Introducing Cyclic Graphs in Neural NetworksDeep Learning is driven by the possibility of computing at-scale gradient of the minimization problem for Directed Acyclic Graphs (DAG). ...

Roman KazinnikJan 14, 20181 minApproximation 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...

Roman KazinnikJan 12, 20187 minBeyond 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...

Roman KazinnikJan 8, 20184 minAn Insight into Data Quality: More Data Equals More Work, But Not Necessarily a Better ModelIn the field of data quality and predictive modeling, which modeling metrics are best at “telling the truth,” so to speak? How did I come...