Please enable JavaScript.
Coggle requires JavaScript to display documents.
Intermediate Machine Learning - Coggle Diagram
Intermediate Machine Learning
Reinforcement Learning
on policy
actor-critic
reward signal
policy gradient
loss function
policy iteration
evaluation
improvement
off policy
automatic differentiation
Learning taking place when expectations are violated
q learning
deep q learning
bellman equation
value function
reward signal
agent, state, action
Sequence/Language Models
classical models
RNNs
vanishing gradients problem
LSTMs and GRUs
HMMs
HMM math
HMM estimation
Kalman Filters
LLMs
compressed internet
thinking slower
bigger is better
factored probebailities
Sequence-Sequence Models
encoder and decoder
bottleneck problem
GRUs and LSTMs
attention
self attention
transformers
positional encoding
keys, queries, values
regression
non-parametric regression
kernels
smoothing kernels
cross validation
bias and variance calculations
mercer kernels
neural tangent kernel
inner product
norm
RKHS
representer thm
basis functions
curse of multidimensionality
kernel density estimator
bias/variance tradeoff
double descent
risk
ridge regression
min norm solution
OLS
leave one out cross validation
lasso
non-parametric bayes
bayes thm
ising model
variational methods
variational autoencoders
gibbs sampler
the ELBO
gaussian processes
graphs
whence equivariance
graph neural networks
discrete graphical models
parallel lasso for discrete models
gibbs sampling and variational infernece
undirected and directed
precision matrix
global markov property
estimation
parallel lasso
graphical lasso (glasso)
scaling behaviors
graph laplacian
neural nets
convolutional neural nets
pooling
tensors
layers
dropout
dense layers
kernel calculations
neural tangent kernel
non-linearity activation functions
backpropagation
random features model