From 00ccb7a22366144caa8278b72f62ea2b5f331d8e Mon Sep 17 00:00:00 2001 From: =?utf8?q?Fran=C3=A7ois=20Fleuret?= Date: Thu, 25 Apr 2024 08:29:45 +0200 Subject: [PATCH] Update. --- dlscore.tex | 163 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 163 insertions(+) create mode 100644 dlscore.tex diff --git a/dlscore.tex b/dlscore.tex new file mode 100644 index 0000000..b72c511 --- /dev/null +++ b/dlscore.tex @@ -0,0 +1,163 @@ +%% -*- mode: latex; mode: reftex; mode: flyspell; coding: utf-8; tex-command: "pdflatex.sh" -*- + +\documentclass[11pt,a4paper,twocolumn,twoside]{article} +\usepackage[a4paper,top=2.5cm,bottom=2cm,left=2.5cm,right=2.5cm]{geometry} +\usepackage[utf8]{inputenc} +\usepackage{cmbright} + +\begin{document} + +\noindent One point per item if you know precisely the meaning of the +listed word(s) + +\section{Machine Learning} + +\begin{enumerate} + + \item VC dimension + \item over-fitting, under-fitting + \item logistic regression + \item Q-value + \item kernel trick + \item boosting + \item feature design + \item linear regression + \item expectation-maximization, GMM + \item SVM + \item Bellman equation + \item decision tree + \item train/validation/test sets + \item naive Bayesian model + \item autoregressive model + \item bias-variance dilemma + \item policy gradient + \item random forest + \item k-NN + \item perceptron algorithm + +\end{enumerate} + + +\section{Deep-Learning} + +\begin{enumerate} + + \item Adam + \item softmax + \item residual connections + \item autograd + \item ReLU + \item dropout + \item CLIP + \item Xavier's initialisation + \item Vanishing gradient + \item LeNet + \item ViT + \item transposed convolution layer + \item checkpoint (during the forward pass) + \item minibatch + \item masked model + \item supervised / unsupervised + \item data augmentation + \item attention block + \item SGD + \item batchnorm + \item gradient clipping + \item tokenizer + \item VAE + \item weight decay + \item GELU + \item LSTM, GRU + \item GAN + \item resnet + \item straight-through estimator + \item convolution layer + \item pre-training / fine-tuning + \item perplexity + \item logits + \item cls token + \item forward pass + \item Transformer (original one), GPT + \item backward pass + \item autoencoder, denoising autoencoder + \item layer norm + \item GNN + \item diffusion model + \item cross-entropy + \item max pooling, average pooling + \item RNN + \item contrastive loss + \item positional encoding + \item causal model + \item attention layer + \item SSL + \item MSE + \item positional encoding + \item tensor + +\end{enumerate} + +\section{Math} + +\begin{enumerate} + \item Hessian + \item random variable + \item matrix + \item entropy, mutual information + \item dot product + \item mean, variance + \item L2 norm + \item chain rule (differentiation) + \item Fourier transform + \item continuity, Lipschitz continuity + \item chain rule (probability) + \item polynomial + \item Cantor's diagonal argument + \item Jacobian + \item linear operator + \item gradient + \item Bayes' thorem + \item vector + \item joint law, product law + \item Gaussian distribution + \item distribution + \item determinant, rank + \item eigen-decomposition, svd + \item maximum likelihood + \item Central Limit Theorem + +\end{enumerate} + +\section{Compute Science} + +\begin{enumerate} + + \item polymorphism + \item recursion + \item value passed by reference + \item binary search + \item quick sort + \item parallel scan + \item mutability + \item Turing machine + \item FP32 + \item iterator + \item interpreter, compiler + \item anonymous function + \item set + \item binary heap + \item mutex + \item cache memory + \item scope of a variable or function + \item dynamic programming + \item hash table + \item big-O notation + \item Turing complete + \item class inheritance + \item closure + \item loop unrolling + \item complexity + +\end{enumerate} + +\end{document} -- 2.39.5