Posts by Year

On HackerNews there was a topic discussed at some point about ways to distil knowledge from a complex (almost black box) large tree ensemble (a RandomForest ...

Converting a DecisionTree into python code

5 minute read

It is sometimes useful to be able to convert a decision tree into and actual useful code snippet. This notebook shows how you’d go around to achieve this.

RetinaFace

6 minute read

#ml #dl #face #detection #image #processing

Bloom’s effect on learning

less than 1 minute read

#learning #teaching

3D Morphable Models

5 minute read

#dl #ml #computer #vision #image #processing #face #recognition #deep-fakes

Replicating_a_famous_optmisation_gif

18 minute read

There is this chart from Alec Radford that became quite famous for showing a comparative run between different optimizers.

Teacher forcing

3 minute read

Training an RNN with teacher forcing. 20200608182759

Optimisation playground

45 minute read

Test, test, test: optimizing the pooling test strategy

7 minute read

Building an ai mechanical nose

8 minute read

Working with cool projects

Machine learning tools

less than 1 minute read

This is an opinionated presentation where I talk about what are the beginner tools that you need to know, when you want to start doing machine learning.

On cross entropy

15 minute read

Cross entropy can be used to define a loss function in machine learning and is usually used when training a classification problem.

Romanian wiki corpus

3 minute read

Tutorial on machine translation

39 minute read

import pandas as pd import numpy as np

Presidential_elections_visualisations

11 minute read

As I’ve said last time we’re going to do some visualization exercises by using matplotlib animation and some custom made charts.

Mapping

6 minute read

Ssa (singular spectrum analisys)

34 minute read

Election analisys (i)

28 minute read

Getting data of an election

23 minute read

Getting data of an election

Cluj ml enthusiasts

11 minute read

My last post received quite some responses, but before actually settling on the curricula of the upcoming ML course, I’ve decided to “eat my own lunch” and u...