Distributed Stochastic Gradient Langevin Dynamics

A Distributed Stochastic Gradient Langevin Dynamics is a stochastic gradient langevin dynamics algorithm that samples minibatches or any uniform random partitioned dataset.

• AKA: D-SGLD.
• Context:
• Example(s):
  • Stochastic Gradient Langevin Dynamics Algorithm
• Counter-Example(s):
  • Stochastic Gradient Descent Algorithm.
• See: Stochastic Gradient Langevin Dynamics Algorithm, Stochastic Gradient Descent Algorithm----

---

References

2014

• (Ahn et al., 2014) ⇒ Sungjin Ahn, Babak Shahbaba and Max Welling (2014)."Distributed Stochastic Gradient MCMC"

---