I'm learning NLP in public, one note at a time.

UBC · Overview of NLP and Large Language Models9 stops · course complete

1. 1 start with the course UBC's Overview of NLP and Large Language Models: the Whole Course at a Glance The whole UBC course in one place: the micro-certificate it belongs to (and the StrongerBC grant), how it runs, the instructor, the course site, and how four modules became eight notes.
2. 2 new to NLP? start here NLP: from language to research applications Words, context, and vocal cues: an introduction to natural language processing and its connection to interview research.
3. 3   Basic Text Processing: From Word Structure to Pattern Matching Connect word structure, normalization and regular expressions to understand how text becomes material for analysis.
4. 4   Advanced Text Processing: Tokens, Sentences, and Labels From tokens and sentence boundaries to POS tagging and named entity recognition: how text is split and labeled. a paper on this pathLearning the natural history of human disease with generative transformers
5. 5   From Word Meaning to Knowledge Graphs: Lexemes, WordNet, FrameNet, and Knowledge Graphs From lexemes and word relations through WordNet and FrameNet: how computers build up to knowledge graphs, and what UMLS and MeSH do in medicine.
6. 6   The Basis of Language Models: Probability From letter and word frequencies through conditional probability and bigrams: how a simple language model scores a sentence and writes one word at a time.
7. 7   Neural Language Models: From Word Vectors to Backpropagation From what every language model computes through embeddings, neurons, activation functions and backpropagation: how a neural network learns to predict the next word.
8. 8   Large Language Models: Tokens, Attention, Prompting and Reasoning From three improvements on neural language models (tokens, longer context and attention) through the Transformer and the GPT and Llama families: prompting, in-context learning, chain-of-thought and models trained to reason.
9. 9   Open Issues in Large Language Models: Hallucination, Bias, Alignment and Interpretability From hallucination and stale information through FActScore, DoveScore and RAG to RLHF, bias mitigation and interpretability: where the problems of large language models come from and how they can be reduced.

UBC · Extraction and Analysis of Clinical Documents3 stops so far

1. 1 start with the course UBC's Extraction and Analysis of Clinical Documents: Course Intro UBC's second NLP course at a glance: where it fits in the micro-certificate, how it runs, who presents it, its four modules, and my first impression after module one.
2. 2 Where Can NLP Help in Health Care? Clinical Documents, Patient Voices and the Cancer Journey Why NLP matters in health care: clinical documents versus patient-written text, four groups of NLP tasks, other languages and speech, and how a cancer journey leaves siloed reports whose fields, and negations, a computer must read.
3. 3 NLP at the BC Cancer Registry: From Pathology Reports to Privacy How the BC Cancer Registry reads a million pathology reports a year: why keyword rules fail on "no evidence of cancer", a pipeline that uses regex, small models and an LLM each where it fits, and how models can leak patient data unless it is anonymized first.
4. 4 Alcohol in the Emergency Department
   · coming next ·Next stopMy notes on Uncovering the Burden of Alcohol in the Emergency Department with NLP

Stanford · Probability for Artificial Intelligence3 stops so far

1. 1 start with the course Stanford's Probability for AI: an Introduction to the Course A free Stanford course on the probability behind modern AI: who teaches it, the one-teacher-per-ten-students model, the weekly section, and the six-week plan.
2. 2 Axioms of Probability: Complement, OR and Independence From probability symbols and how to read them through Kolmogorov's three axioms, the complement and OR rules and conditional probability to independence: when two probabilities can simply be multiplied, and when conditioning changes that.
3. 3 Random Variables and Three Common Distributions: From the PMF to AI Watermarks From random variables and the probability mass function through expectation and variance to the Bernoulli, binomial and Poisson distributions, and how changing word probabilities lets an AI text carry a watermark.
4. 4 Probabilistic Models
   · coming next ·Next stopMy notes from week two, lesson 2 of Probability for AI