All Articles

Extended archive of working documents, literature reviews, and code snippets. These notes reflect my ongoing learning process.

Analiza tego, jak dwuwarstwowy perceptron uczy się reprezentować dodawanie modulo p poprzez teorię charakterów grup cyklicznych.

Comparing five gradient-based optimization methods across three synthetic non-convex landscapes.

Analysis of self-attention mechanisms.

Step-by-step mathematical derivation of the MLE for the mean vector and covariance matrix.

Formal derivation and error analysis of the Euler method, developed as an extension of an optional discretization task.