Short (More Technical) Notes

(referred to in the main articles, but not essential to reading them)


What is a computation or an algorithm? The Church-Turing thesis posits that the equivalent formulations of computation (e.g. a Turing machine) are providing the answer.

The World of the Neuron

The basic algorithms used in deep learning, the so-called artificial neural networks, have borrowed the concept of neuron and synaptic activation from the biological neuron.

Truth and Provability

The relationship between truth and provability (and between formal systems and their models) is fundamental to understanding the foundational questions in AI.

An introduction to denotational semantics, namely that the meaning of words is mainly contextual. The modern treatment of this denotational semantics approach is based on mapping each word to a vector.


Neural networks have awesome performance for fact verification, but do it only in a black-box fashion, without explainability. We need additional tools because humans need explanations in natural language.


Transformers have been the most interesting new development in neural networks, especially in language applications. We will try to understand how they are built and what their range of applications is.