Autoregressive model
Also: AR model
An autoregressive model factorizes the joint probability of a sequence with the chain rule into a product of next-element conditionals. GPT-style language models, PixelCNN and WaveNet all work this way; likelihood is exact, but generation is sequential.
P(x₁, …, xₙ) = Πₜ P(xₜ | x₁, …, xₜ₋₁)
Example
If P("the") = 0.05, P("cat" | "the") = 0.01 and P("sat" | "the cat") = 0.2, then P("the cat sat") = 0.05 · 0.01 · 0.2 = 0.0001.