Reading Maths Notation in AI Papers
Translate the symbols in machine learning papers into plain words and small examples
A taste of a lesson
A paper writes θ* = argmin_θ (1/n) Σ_i L(f_θ(x_i), y_i). How do I read this?
Read it from the inside out. f_θ(x_i) is the model's prediction for example i, using parameters θ. L compares that prediction with the true answer y_i, giving a loss. Σ_i adds the losses over all examples, and 1/n turns the sum into an average. argmin_θ means: the parameter values that make this average loss as small as possible. θ* names those best parameters. In words: train the model to minimise average loss on the data. Try it: with two examples and losses 0.4 and 0.2, what is the average loss?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Read common symbols for sums, sets, vectors, matrices and norms aloud
- Interpret expectations, conditional probabilities and distributions in notation
- Decode gradients, partial derivatives and argmin expressions
- Apply a step by step method to any unfamiliar equation
- Check understanding with tiny numeric examples and shapes
Lesson plan
- 1 Sums, products and indices Read summation and product notation fluently. Start
- 2 Sets, vectors and matrices Interpret membership, spaces and shapes. Start
- 3 Norms, dot products and transposes Read length and similarity notation. Start
- 4 Probability notation Decode conditional probabilities, distributions and expectations. Start
- 5 Calculus and optimisation notation Read gradients, partial derivatives and argmin. Start
- 6 Decoding a real equation Apply a full method to equations from papers. Start
Try asking
About this tutor
A tutor for any level that helps you read the mathematical notation in AI papers, documentation and textbooks without freezing. You will learn the common symbols for sums, products, sets, vectors, matrices, norms, expectations, conditional probabilities, gradients and optimisation, and a reliable method for decoding an unfamiliar equation: name every symbol, read it aloud in words, try a tiny numeric example and, if useful, write it as a few lines of pseudocode. Bring equations from papers you are reading and the tutor will work through them with you at your pace.
Reviews
4.7
3 ratingsSample
- Patrick O.Sample
Very patient. It honestly flagged where a paper's notation was ambiguous, which was reassuring.
- Carmen V.Sample
As a software engineer moving into ML, this removed my fear of papers. The pseudocode translations were perfect for me.
- Hiroshi N.Sample
I bring an equation from each paper I read and we decode it together. The two point example trick works every time.
About the teacher
Linear algebra for AI, with geometry first and notation second
9 tutors 338 lessons taught Sample
I teach the linear algebra behind modern AI: vectors, matrices, similarity, eigenvectors and the methods built on them, such as PCA, clustering and recommender systems. I trained in applied mathematics and later worked on search and recommendation features, so I like to connect each idea to something a real system does. My lessons begin with pictures and small numbers you...
See Katarzyna's profile and tutorsMore like this
Other tutors on the same or nearby topics.