[Virtual] PyData Eindhoven Meetup

Are you missing your friends from the Python community? PyData Eindhoven is back with a new virtual meetup.

Agenda

• 16:00 - Short Introduction
• 16:05 - Measuring micron-sized particles with light and machine learning by Alican Noyan the founder of Ipsumo
• 16:40 - Break (using Mibo)
• 17:10 - A Gentle Introduction To Multi Objective Optimisation by Eyal Kazin - Senior Data Scientist from Babylon Health
• 17:40 - Questions/Virtual Drinks (in Mibo)
  ---
  As a side note if you live in the Netherlands we will send some popcorn to your home for free. If you would like popcorn sent to you fill out this form: https://forms.gle/9W2eSM7QmebyQPVLA
  ---
  Title: Measuring micron-sized particles with light and machine learning
  Short Description: Neglect friction, assume constant temperature, consider a spherical cow... Physics-based models rely on assumptions to reduce complexity, but what happens when those assumptions are violated? In this talk, I will explain how I helped an international team of physicists and engineers build a novel particle size measurement device using machine learning. Their design provided significant size and cost reduction over traditional systems. Yet the underlying physics theory assumed single-scattering whereas in reality light scattered from particles more than once. This led to an underestimation of the particle size. This is the story of a random forest model that learned how to make accurate particle size predictions from light scattering data.
  This talk will go into more detail around this published article: https://www.nature.com/articles/s41377-020-0255-6
  Presenter: Alican Noyan - Founder of Ipsumio https://www.ipsumio.com/
  LinkedIn Profile: https://www.linkedin.com/in/mehmetalicannoyan/

---
Title: A Gentle Introduction To Multi Objective Optimisation
Short Description:Optimising for multiple objectives is a non-trivial task, especially when they are in conflict. For example how can one best overcome the classic trade-off between quality and cost of production, when the monetary value of quality is not defined? In this 30 minute talk you will learn about Pareto Fronts and how they can be used to optimise for multiple objectives simultaneously using Genetic Algorithms.
This talk is geared towards anyone interested in improving their decision making and optimisation skills (e.g, analysts, scientists, engineers, economists), in which you will learn:

• The limitations of the common practice of combining multiple parameters into one heuristic.
  *Pareto Fronts: the notion in which there may be a set of trade-off solutions which are considered equally optimal.
  *Pareto Fronts in the context of Genetic Algorithms
  *The advantages of prototyping with python module DEAP (https://deap.readthedocs.io/en/master/) .
  Application to therapeutic discovery and machine learning hyperparameter tuning

My objective is for you to get a basic intuition for the technique, understand its advantages and shortcomings to be able to assess applicability for your own projects.

For those interested in gaining hands-on experience, see my tutorial in github (https://bit.ly/github-moo), where you will practice the knapsack problem(https://en.wikipedia.org/wiki/Knapsack_problem). Here you will program for filling a bag with packages with the objective of minimising the bag weight while maximising its content value.

Presenter: Eyal Kazin – Senior Data Scientist at Babylon Health (London)
LinkedIn Profile: https://www.linkedin.com/in/eyal-kazin-0b96227a/

Attendee profile:

• People who want to learn about Scientific Computing
• Opensource enthusiasts
• Interested in seeing how others use Python/R/Julia
• Interested in being part of the PyData Community in Eindhoven
• No Recruiters
• No Sales