GFlowNets cast sampling from an unnormalized discrete distribution as a flow-conservation problem, and the thesis applies them to approximate Bayesian posteriors over DAGs.
Proposition A.1.2 (Data processing inequality)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Generative Flow Networks: Theory and Applications to Structure Learning
GFlowNets cast sampling from an unnormalized discrete distribution as a flow-conservation problem, and the thesis applies them to approximate Bayesian posteriors over DAGs.