This month we delve into quantum mechanics and learn about and how Deepmind researcher Sander and his team won the Kaggle plankton classification competition.
Skillsmatter link: : https://skillsmatter.com/meetups/7312-classifying-plankton
See you there!
• Algorithms that crack quantum mechanics: from atoms to real materials - Mattia Fiorentini (http://twitter.com/matt_fiorentini)
Since the rise of nanotechnology increasing efforts have been made to develop material-design strategies directly from the first principles of matter. In particular, the properties of any materials, even the ones that have not been synthesized yet, can be predicted by solving the equations of quantum mechanics. In this talk, I will show how this highly non-trivial task can be pursued using two well known algorithm such as the Fourier transform and the conjugate gradient, in combination with powerful computer clusters. I will also present how this computational techniques can be beneficial to the development of more efficient thermoelectric materials.
Bio: I am computational physicist and PhD candidate at King's College London. My main efforts are currently devoted to development of an HPC software framework to study thermoelectric phenomena in real materials from first-principles. I am also also active in scientific dissemination organizing the Pint of Science festival. Data science, and finance are among my side interests.
• Classifying plankton with deep neural networks - Sander Dieleman (http://twitter.com/sedielem)
Deep neural networks have become very popular for solving computer vision problems in recent years. This talk is an overview of a practical application of this approach: I'll show how our team of 7 built a model for the automated classification of plankton based on convolutional neural networks. Using this model, we placed 1st in the National Data Science Bowl competition on Kaggle.
Bio: Sander Dieleman is a research scientist at Google DeepMind and a PhD student in the Reservoir Lab at Ghent University in Belgium. The main focus of his PhD research is applying deep learning and feature learning techniques to music information retrieval (MIR) problems, such as audio-based music classification, automatic tagging and music recommendation.