In school we round numbers like 0.5, 1123.5 towards the bigger number. It’s a “round half up” method.
That introduces an undesired bias some cases. For example if we have a large data set, and we aggregate some column containing a lot of .5 fractions. In order to adjust for it in many cases a rounding of 0.5 towards nearest even number is applied. It’s “Rounding half to even” or “banker’s rounding”....
With the help of python and SymPy module one can do pretty neat computations. For example when I took a course about Robotic Preception on Coursera I had to find a cross product of two vectors v1 x v2 represented in a generic form:
v1 = (a, b, c) v2 = (d, e, 0) Normally I would write it down on a piece of paper and do the computations myself. Luckily python can help with that....
Here I will collect links to well-written articles about math and algorithms.
Visual Information Theory. Nice post about probabilities, entropy. The whole blog has awesome visualizations. Highly recommended.
MNIST For ML Beginners - Nice introduction into machine learning with TensorFlow
Brilliantly wrong - blog by Alex Rogozhnikov, with fantastic visualizations of many algorithms and concepts
Collection of data science notebooks.
Deconvolution and Checkerboard Artifacts by Odena, et al., 2016 - explanation of deconvolution operation in neural networks with awesome interactive visualizations...