Profile cover photo
Profile photo
Kasper Fredenslund
76 followers
76 followers
About
Posts

Post has attachment
Photo
Add a comment...

Post has attachment
Photo
Add a comment...

Post has attachment
Photo
Photo
5/20/18
2 Photos - View album
Add a comment...

Post has attachment
Photo
Photo
5/20/18
2 Photos - View album
Add a comment...

Post has attachment
Introduction To Neural Networks

In recent years, neural networks have shown great potential across a wide range of industries. In this series, we look at how neural networks work from a theoretical point of view.

https://kasperfred.com/series/introduction-to-neural-networks
Add a comment...

Post has attachment
Introduction to computational complexity.

A look at different ways of determining the efficiency of algorithms.

From Big-O to empirical measurements, the series interweaves theory and practice to provide a strong overview of the most important ideas.

https://kasperfred.com/series/computational-complexity
Add a comment...

Post has attachment
Post Mortem Analysis of my Final Year Project

A post mortem analysis of a Data Science approach for determining the existence and decay patterns of the Higgs boson.


In 2013, the CERN LHC Atlas team released a dataset containing simulated proton-proton collisions some of which resulted in a 125 GeV Higgs which would decay to two taus. Others resulted in two top quarks decaying to either a lepton or a tau, or a W boson decaying to either an electron or a muon and tau pair.

The problem was construct a neural network that correctly segmentize the events using features that can either be measured directly in the accelerator, or can be derived from the measurements.

The final network consisting of only a single hidden layer was able to predict the existence of the Higgs boson with an accuracy of 99.997% over 5 positive predictions (the network thinking that the Higgs exists). For such a simple network, I find that number to be rather impressive.

This piece will be a summarization of my immediate reflections following the project....

https://kasperfred.com/posts/post-mortem-analysis-of-my-final-year-project
Add a comment...

Post has attachment
A practical application of convolutional neural networks: Looking at traffic signs.

You know, I don't think we as a species do enough looking at German traffic signs.

I mean, sure, they are there when we drive through Germany, and we do (hopefully) see them, and sometimes we even register their meaning, and alter our behavior based on those meanings. But we don't do nearly enough looking at those bold, blue, red, and white, geometrical pictograms.

I think this is a shame, as by virtue of not looking, we do not appreciate their simplistic genius of interlingual communication.

That's why I decided to outsource all the looking to computers a while back when I was toying with an image classifier that would learn to classify traffic signs...

https://kasperfred.com/posts/looking-at-german-traffic-signs
Add a comment...

Post has attachment
Introduction to Tensorflow as a Computational Framework

Tensorflow is likely the most popular, and fastest growing machine learning framework that exists. With over 70000 stars on Github, and backing from Google, it not only has more stars than Linux, but also has a ton of resources behind it.

If that doesn't peak your interest, I have no idea what will.

In this tutorial we will talk about:

- General design philosophy
- Visualization
- Examples covering common use cases
- How it relates to machine learning

https://kasperfred.com/posts/introduction-to-tensorflow-as-a-computational-library/
Add a comment...

Post has attachment

You know these internet memes that say if you can solve the question they ask, you are a certified genius? They seem to creep up on you no matter what you do to stop it.

Often, the answer is obvious, however, I'd much rather expend my energy watching cat videos than proving my genius to everyone.

This is exactly what we going to be doing today.

By the end of this tutorial you will be able to:

- Use various regression algorithms to solve internet genius memes.
- Understand how different algorithms can capture different relationships in the data.
- Using grid search to optimize multiple algorithms across a range of different parameters automatically.

https://kasperfred.com/posts/solving-only-1-can-answer-this-problems-with-machine-learning/
Add a comment...
Wait while more posts are being loaded