Presentation of PolyMath at ESUG 2022 in Novi Sad, Serbia

1. version 1.0 v2.0
Oleksandr Zaitsev, Sebastian Jordan Montaño,
Serge Stinckwich, Hemal Varambhia, etc.

2. 2
PolyMath is a library for
scientific computing in Pharo.
(it’s a math library)
In a nutshell:

3. 3
On ESUG 2019 in Cologne,
Serge Stinckwich presented PolyMath v1.0

4. Time flies :)
2019 2020 2021 2022
…

5. 5
Now we are here

6. 6
1. Remind you what is PolyMath
2. Discuss what lies ahead
3. Tell you how to support us
So today we will:

7. Part 1:
PolyMath -
Who are we?

8. 8
We are not mathematicians

9. 9
Contributors (between v1.0 and now)
Serge
Stinckwich
Hemal
Varambhia
Oleksandr
Zaitsev
Yvan
Guifo
Sebastian
Jordan Montaño
Hernan
Moralez Durand
Rakshit

10. SVD
10
Some Packages & Algorithms
Complex
Numbers
Probability
Distributions
Kolmogorov
Smirnov
Matrix
Decomposition
Dimensionality
Reduction
Differential
Equations
Quaternion
Statistics
Polynomials
Numerical
Methods
Matrix-Vector
Algebra
Numerical
Optimisation
DataFrame

11. 11
Getting Started with PolyMath
Metacello new
repository: 'github://PolyMathOrg/PolyMath';
baseline: 'PolyMath';
load.

12. 12
Two Data Structures
Vector Matrix
vector := PMVector withAll:
#(1 4 3 -1).
matrix := PMMatrix rows: #(
(1 0 0)
(0 1 0)
(0 0 1)).

13. 13
Dimensionality Reduction with PCA

14. 14
Dimensionality Reduction with PCA
- Used in machine learning, statistics to solve the
dimensionality curse.
- Dimensionality reduction approach that perform a
linear mapping of the data to a lower-dimensional
space, in such a way that the variance of the data in
low-dimensional representation is maximised.

15. 15
Dimensionality Reduction with PCA

16. 16
Dimensionality Reduction with PCA

17. 17
Dimensionality Reduction with PCA
Visualised
with Roassal3
❤

18. 18
DataFrame

19. 19
DataFrame

20. 20
DataFrame

21. 21
PolyMath Book
“Object-Oriented
Implementation of Numerical
Methods. An Introduction with
Java and Smalltalk”
by Didier B. Besset

22. Part 2:
The future

23. 23
Goal 1: Decoupling Packages

24. 24
Goal 2: Making PolyMath Faster
Time in seconds (less is better)
Small
Medium
Pure Pharo vs Pharo & LAPACK
1820 times faster
2103 times faster
Pharo Pharo & LAPACK
19min40sec 1h44min
0.6sec 2.9sec
Help

25. 25
Goal 3: More Integration with pharo-ai
LAPACK SciPy scikit-learn
pharo-ai
PolyMath
LAPACK
Pharo
Python
Fast
Algebra
Math
Machine
Learning

26. 26
Goal 4: Document Using Microdown
❤

27. 27
Goal 4: Document Using Microdown
❤

28. 28
Goal 5: Roassal Charting

29. 29
Goal 5: Roassal Charting

30. Goal 6: Notebooks
30

31. Goal 6: Notebooks
31

32. Part 3:
How to
support us

33. 33
Contribute!

34. 34
Sponsor our work
https://github.com/
sponsors/SergeStinckwich