Greetings
Knowledge is a paradox. The more one understand, the more one realises the vastness of his ignorance.
—Viktor (Arcane, Season 2)
You have stepped onto the stickiest page of this site, here sit all of my projects.
The legend is as follows:
- Archived.
- Done; でも、 \(\exists\) room for extension / improvement.
- Editing; code written, ideas fleshed out - prose needs to be reworked.
- Refactor; initial code included, but contains bugs, is incorrect or needs better Design Patterns.
- Nothing. All I have is the seedling of an idea.
Note: The unticked projects do not come with any warranty. You may be mislead; \(\mathfrak{hic\,\,sunt\,\,dracones}\).
You may also view the entire tag taxonomy here.
Project List
Classical Computer Science
-
Memory #notes#stack-pointer#registers#cache#L1#L2#stack#heap
- Computational Complexity #big-oh#np-complete#polynomial-time#np-hard
- Internet Networks #tcp#ip#udp#osi#packets#handshake
- Databases #sql#nosql#time-series
-
Linux #OS#monolithic#structure#kernel#linus torvalds#cache#paging
- Version Control #git
Data Structures & Algorithms
- Peg Solitaire #bfs#dfs#memory#puzzle#combinatorics
- Banagrams Solver #solver#python#tries
- Advanced Algorithms #pagerank#dpc#collaborative-filtering#game-theory#bayesian#networks#internet
- Leetcode #problem-solving#employment
- Sorting #big-oh#stable#online#parallel
- Searching #naive#heuristic
- Graph #acyclic#digraph#max-flow#shortest-path
Machine Learning
I have thought about this ML hierarchy inasmuch as Aristotle thought about the phylums of flowers.
I am not a Data Scientist, but rather a Computer Scientist and Mathematician.
As such, my interests lie in theory giving rise to applications. Not vice-versa–applications giving rise to theory–which I believe retard the habit of generalisation and thus imagination.
Datasets
The following are all tags, but visiting them provides contextual / historical information on the dataset as well as back-links to the models which have solved these problems.
- MNIST #classification#logistic-regression#softmax#lecun#cnn#random-forests
- KMNIST #classification#multiclass#cnn#2-layer
- FMNIST #classification#multiclass#random-forest#mnist
- CIFAR #classification#image#cnn
- IRIS #classification#multiclass#cnn#svm
- ImageNet #classification#resnet-50
- California Housing
- Wine Quality #ensembles#gradient-boosting
- IMDB Reviews #BERT
- Pima Indians Diabetes #decision-tree
- WHO Life Expectancy #WHO
- Titanic Deaths #logistic-regression#binary-classification#classification#kaggle
- KDD Cup 1999 #k-means#clustering
- Digits #gaussian-mixture-model
Theory
- Backpropagation
- Cross Validation #linear-regression#pandas#loocv#lasso#ridge#eda#hyperparameter-tuning
- No Free Lunch Theorem
- Curse of Dimensionality
- Bias Variance Decomposition #overfitting#generalisation
- Performance Metrics #sensitivity#specificity#roc-curve#confusion-matrix
- Loss Functions
- Kernel Methods
Supervised Learning
Classification
These methods can be adapted for regression, but they are more well suited to classification.
Deep Learning
- Hardware Benchmarking #benchmark#ml#dl#macos#m1
- Perceptrons with Gradient Descent (Sigmoid Loss) #gradient-descent#sigmoid
- Multi-layered Perceptron #xor#feedforward-networks#backpropagation#pytorch#hidden-units#perceptron#mnist
- Deep Neural Networks #feedforward#deep#back-prop#grad-descent#regularisation#neural-nets#cnn
- Recurrent Neural Networks (RNN) #sequence-to-sequence
- Long Short-Term Memory (LSTM) #gates
- Convolutional Neural Networks (CNN) #scratch
- Transformers #attention
- Autoencoders
- Generative Adversarial Networks (GAN's) #multimodal
- Stable Diffusion #state-of-the-art
Mathematics
N-Bday Problems
The compiled PDFs can be found in the above linked heading.
The following links contain the source code and solution sets: