Articles and videos — everything I've made to explain something, most recent first.
The output distribution was a choice. Other choices give other costs, each predicting a different statistic — and why cross-entropy beats squared error when the gradient would otherwise fall silent.
Cost functions derived from maximum likelihood. Why probabilities multiply, how the logarithm keeps the order— the top-down route to the loss we assumed in 1.1.
What a neural network is before any of the math: why deep learning had to exist, what every term actually means, and why a model must learn its own features. The one model the whole field grows from, built from zero.
How the 2009 GNN model actually works: the state update and output functions, the two-phase learning algorithm, the training flow, and the linear vs nonlinear transition functions.
The theory behind Scarselli et al.'s original GNN paper: the tau function, information diffusion, universal approximation, and the positional/non-positional distinction.
How a 1951 statistics formula quietly keeps your chatbot from going feral. KL divergence, RLHF, and the tension between them.
The derivative, derived by hand from MSE, and gradient descent built on top of it. Where the 2/n comes from, why we subtract the gradient, and a full worked training example.
Slope, linear regression, and the error function, built from the ground up. Why we divide by n, why we square the error, and why a loss value alone cannot tell you how to improve.
Matplotlib as the engine behind every Python chart, Seaborn for statistical plots, and when to reach for each — with 30+ chart types.
Fixing what is wrong before it breaks your analysis, then combining multiple DataFrames into one clean dataset.
Splitting data into groups and aggregating them, plus label- and position-based indexing with loc and iloc.
Loading data into Pandas from CSV, Excel, JSON, HTML, SQL and more — then sorting and displaying it to understand a dataset before you analyse it.
Single Instruction Multiple Data in Python, explained with a concrete mathematical example — why NumPy vectorizes and lists don't.
Why hypergraphs are the promised representation for learning networks — higher-order relationships, spatial context, and why HyGNN beats plain GCN.
How graph and hypergraph neural networks turned drug interaction prediction from empirical guesswork into a computational discipline — and the cold-start problem still holding it back.
How the State Pattern lets an object change its behavior when its internal state changes — replacing a tangle of conditionals with clean, self-contained state classes, built around a student moving through exam stages.
A structural design pattern that provides a surrogate or placeholder for another object to control access to it.
The time complexity of common operations across arrays, linked lists, and trees — and how to pick the right structure before scale bites you.
How one matrix bridges discrete graphs and continuous math — Laplacian smoothing, spectral analysis, and why GNNs depend on it.
How to generate PDF documents from HTML in an ASP.NET Core app using .NET 8 — a clean six-step build with the open-source DinkToPdf, plus how the same logic looks with the commercial IronPDF.
How can you write a graduation research paper for a bachelor's degree? We present the methodology that the student should follow to build the graduation research file.
How can you write a graduation research paper for a bachelor's degree? We present the methodology that the student should follow to build the graduation research file.
How can you write a graduation research paper for a bachelor's degree? We present the methodology that the student should follow to build the graduation research file.
The distinction between a programmer and a developer, the skills a web programmer needs, and why a strong web programmer is an asset to any IT team.