From Self-Attention to Markov Models: Unveiling the Dynamics of Generative Transformers.
Muhammed Emrullah IldizYixiao HuangYingcong LiAnkit Singh RawatSamet OymakPublished in: CoRR (2024)
Keyphrases
- markov models
- markov model
- maximum entropy
- higher order
- hidden markov models
- low order
- transition probabilities
- sequence classification
- hidden state
- generative model
- conditional random fields
- dynamical models
- sequence prediction
- computer vision
- statistical model
- logic programs
- graphical models
- upper bound
- special case