Meetup Bratislava
This is a group for anyone who wants to explore the area of data science, artificial intelligence, machine learning. All skill
and experience levels are welcome. I started this group to meet other outdoor enthusiasts of areas as:
• build systems that use machine learning techniques in Clojure, Java, Node.js
• understand machine learning problems such as regression, classification, and clustering
• discover the data structures used in machine learning techniques such as artificial neural networks and support vector
machines
• implement machine learning algorithms in real scientific or business cases
• learn more about software libraries to build machine learning systems
• discover techniques to improve and debug solutions built on machine learning techniques
• use machine learning techniques in a cloud architecture for the modern Web
branislav.majernik@oracle.com
DataData
Meetup Bratislava
Vážení priaznivci moderného IT,
Prijmite naše pozvanie na dobrú kávu s crossisantom v priestoroch kaviarne KORZO na Hviezdoslavovom námestí 3 v Starom
Meste, dňa 14.2.2018 od 9:00 – 11:00. Naštartujeme deň diskusiou o aktuálnych trendoch v informačných technológiách,
o data science, machine learning a IT s tým súvisiacim.
• Support Vector Machine – prečo je úspešná v segmentácii komplikovaného zákazníka
• Markov chains a ich využitie v predikcii správania sa zákazníka
Ak ste sa v téme nenašli, nezúfajte, nasledovať bude:
Lambda Coffee - venovaný funkcionálnej paradigme a technológiám.
Registrácia na adrese:
branislav.majernik@oracle.com
klaudia.blaskovicova@oracle.com
DataData
What is data science ?
What is data science?
Tasks of regression, interpolation
What is data science?
Tasks of classification
What is data science?
How to learn computer a new function without explicit programming
Data science needs
Platform
Mathematical
Native Language Speed
Data Structures
Language Constructs
Data Connectivity
End-2-End Quality Control
Interactive Environment
Visualization
Linear Algebra
Statistics
Machine Learning
Optimizers
R, Mathlab, Mathematica
C++, Java, Python
Excel
Data science needs
 Productive and fun
 Portability, good parts of JVM ( )
 REPL – interactive experiments
 Functional programing
 DSL's with composable abstractions
 Data is code, code is data
and deployment in ...
spread of platform,
on the fly profiling,
inlining, loop-unrolling,
de-opt/reopt,
escape analysis,
dead code elimination,
proven GC
Support Vector Machine
Tasks of classification
Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
Support Vector Machine
Tasks of classification
Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
Support Vector Machine
Tasks of classification
Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
Support Vector Machine
Tasks of classification
Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
Support Vector Machine
Tasks of classification and regression
Support Vector Machine
Tasks of classification and regression
Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
Support Vector Machine
Tasks of classification and regression
Source of this picture: http://web.mit.edu/6.034/wwwbob/svm-notes-long-08.pdf
Support Vector Machine
Tasks of classification and regression
Source of this picture: http://web.mit.edu/6.034/wwwbob/svm-notes-long-08.pdf
Support Vector Machine
Tasks of classification and regression
Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
Support Vector Machine
Tasks of classification
Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
Support Vector Machine
Tasks of classification
Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
Support Vector Machine
Tasks of classification
Source of this picture: http://scikit-learn.org/stable/_images/sphx_glr_plot_rbf_parameters_001.png
Support Vector Machine
Tasks of regression
Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
Support Vector Machine
Tasks of regression
Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
Support Vector Machine
Tasks of regression
Source of this picture: https://www.svm-tutorial.com/2014/10/support-vector-regression-r/
Support Vector Machine
Correspondence with Radial Basis Neural Network
Source of this picture: http://mccormickml.com/2013/08/15/radial-basis-function-network-rbfn-tutorial/
Support Vector Machine
With genetic programing evolution
Source of this picture: https://github.com/ssusnic/Machine-Learning-Flappy-Bird
Markov chain
When probability depend on previous outcome
Hidden Markov chain
When probability depend on previous outcome and is hidden from observable state
Hidden Markov chain
When probability depend on previous outcome and is hidden from observable state
calculus
branislav.majernik@oracle.com
Lambda
calculus
branislav.majernik@oracle.com
Lambda

Data coffee - Support vector machine usage with complex data

  • 1.
    Meetup Bratislava This isa group for anyone who wants to explore the area of data science, artificial intelligence, machine learning. All skill and experience levels are welcome. I started this group to meet other outdoor enthusiasts of areas as: • build systems that use machine learning techniques in Clojure, Java, Node.js • understand machine learning problems such as regression, classification, and clustering • discover the data structures used in machine learning techniques such as artificial neural networks and support vector machines • implement machine learning algorithms in real scientific or business cases • learn more about software libraries to build machine learning systems • discover techniques to improve and debug solutions built on machine learning techniques • use machine learning techniques in a cloud architecture for the modern Web branislav.majernik@oracle.com DataData
  • 2.
    Meetup Bratislava Vážení priaznivcimoderného IT, Prijmite naše pozvanie na dobrú kávu s crossisantom v priestoroch kaviarne KORZO na Hviezdoslavovom námestí 3 v Starom Meste, dňa 14.2.2018 od 9:00 – 11:00. Naštartujeme deň diskusiou o aktuálnych trendoch v informačných technológiách, o data science, machine learning a IT s tým súvisiacim. • Support Vector Machine – prečo je úspešná v segmentácii komplikovaného zákazníka • Markov chains a ich využitie v predikcii správania sa zákazníka Ak ste sa v téme nenašli, nezúfajte, nasledovať bude: Lambda Coffee - venovaný funkcionálnej paradigme a technológiám. Registrácia na adrese: branislav.majernik@oracle.com klaudia.blaskovicova@oracle.com DataData
  • 3.
    What is datascience ?
  • 4.
    What is datascience? Tasks of regression, interpolation
  • 5.
    What is datascience? Tasks of classification
  • 6.
    What is datascience? How to learn computer a new function without explicit programming
  • 7.
    Data science needs Platform Mathematical NativeLanguage Speed Data Structures Language Constructs Data Connectivity End-2-End Quality Control Interactive Environment Visualization Linear Algebra Statistics Machine Learning Optimizers R, Mathlab, Mathematica C++, Java, Python Excel
  • 8.
    Data science needs Productive and fun  Portability, good parts of JVM ( )  REPL – interactive experiments  Functional programing  DSL's with composable abstractions  Data is code, code is data and deployment in ... spread of platform, on the fly profiling, inlining, loop-unrolling, de-opt/reopt, escape analysis, dead code elimination, proven GC
  • 9.
    Support Vector Machine Tasksof classification Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
  • 10.
    Support Vector Machine Tasksof classification Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
  • 11.
    Support Vector Machine Tasksof classification Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
  • 12.
    Support Vector Machine Tasksof classification Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
  • 13.
    Support Vector Machine Tasksof classification and regression
  • 14.
    Support Vector Machine Tasksof classification and regression Source of this picture: https://med.nyu.edu/chibi/sites/default/files/chibi/Final.pdf
  • 15.
    Support Vector Machine Tasksof classification and regression Source of this picture: http://web.mit.edu/6.034/wwwbob/svm-notes-long-08.pdf
  • 16.
    Support Vector Machine Tasksof classification and regression Source of this picture: http://web.mit.edu/6.034/wwwbob/svm-notes-long-08.pdf
  • 17.
    Support Vector Machine Tasksof classification and regression Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
  • 18.
    Support Vector Machine Tasksof classification Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
  • 19.
    Support Vector Machine Tasksof classification Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
  • 20.
    Support Vector Machine Tasksof classification Source of this picture: http://scikit-learn.org/stable/_images/sphx_glr_plot_rbf_parameters_001.png
  • 21.
    Support Vector Machine Tasksof regression Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
  • 22.
    Support Vector Machine Tasksof regression Source of this picture: http://www.robots.ox.ac.uk/~az/lectures/ml/lect3.pdf
  • 23.
    Support Vector Machine Tasksof regression Source of this picture: https://www.svm-tutorial.com/2014/10/support-vector-regression-r/
  • 24.
    Support Vector Machine Correspondencewith Radial Basis Neural Network Source of this picture: http://mccormickml.com/2013/08/15/radial-basis-function-network-rbfn-tutorial/
  • 25.
    Support Vector Machine Withgenetic programing evolution Source of this picture: https://github.com/ssusnic/Machine-Learning-Flappy-Bird
  • 26.
    Markov chain When probabilitydepend on previous outcome
  • 27.
    Hidden Markov chain Whenprobability depend on previous outcome and is hidden from observable state
  • 28.
    Hidden Markov chain Whenprobability depend on previous outcome and is hidden from observable state
  • 29.
  • 30.