Bounding the Error of Discretized Langevin Algorithms for Non-Strongly Log-Concave Targets.
Arnak S. DalalyanAvetik G. KaragulyanLionel Riou-DurandPublished in: J. Mach. Learn. Res. (2022)
Keyphrases
- computational cost
- learning algorithm
- error bounds
- recently developed
- theoretical analysis
- computational efficiency
- information systems
- data sets
- computationally efficient
- graph theory
- convergence rate
- times faster
- computationally expensive
- orders of magnitude
- error rate
- optimization problems
- worst case
- significant improvement
- lower bound
- objective function
- social networks
- real world