Transformer
Attention, sequence modeling, and why the transformer became such a decisive shift in modern NLP.
Open note →Ideas
Natural Language Processing
My notes and takeaways on natural language processing
I wanted this section to be a place where I can organize my NLP notes more carefully instead of keeping everything on one long page. Some of these entries are technical summaries, some are concept notes, and some are simply my way of understanding the foundations more clearly by writing them out.
Right now, I am especially interested in the representational side of NLP: how words become vectors, how context changes meaning, and how architectural shifts like the transformer changed what language models can learn at scale. I also like tracing how older ideas still quietly sit underneath newer ones. Even when the field moves fast, the basic conceptual questions remain surprisingly stable.
In this section
A central map for the sequence, so each note can stand alone while still returning to the question I keep circling.
Neural Language Models and Word Representation
From one-hot vectors to embeddings, CBOW, and Skip-gram: a note on how representation learning changed NLP.
Open note →Model Evaluation
A note on the challenges of evaluating language models, from perplexity to human evaluation and everything in between.
Open note →Parsing
A note on syntactic parsing, from early rule-based systems to modern neural approaches and their implications for understanding language structure.
Open note →