duality gap
Constrained Reinforcement Learning Has Zero Duality Gap
Santiago Paternain, Luiz Chamon, Miguel Calvo-Fullana, Alejandro Ribeiro
Neural Information Processing Systems http://nips.cc/
- South America > Chile > Santiago Metropolitan Region > Santiago Province > Santiago (0.04)
- North America > United States > Pennsylvania (0.04)
- North America > Canada (0.04)
- Europe > United Kingdom > England > Cambridgeshire > Cambridge (0.04)
Primal-DualBlockGeneralized Frank-Wolfe
We propose a generalized variant of Frank-Wolfe algorithm for solving a class ofsparse/low-rank optimization problems.
Uncoupled and Convergent Learning in Two-Player Zero-Sum Markov Games with Bandit Feedback
Our paper takes the first step to remove such assumptions.
- North America > United States > California (0.14)
- North America > United States > Virginia (0.04)
- Europe > United Kingdom > England > Cambridgeshire > Cambridge (0.04)
Hamiltonian descent for composite objectives
Brendan O'Donoghue, Chris J. Maddison
Neural Information Processing Systems http://nips.cc/
- Asia > Middle East > Jordan (0.04)
- North America > Canada > Ontario > Hamilton (0.04)
- Europe > United Kingdom > England > Oxfordshire > Oxford (0.04)
- (3 more...)
- North America > Canada > Quebec (0.04)
- North America > Canada > British Columbia (0.04)
- Europe > France > Auvergne-Rhône-Alpes > Lyon > Lyon (0.04)
A Domain Agnostic Measure for Monitoring and Evaluating GANs
Paulina Grnarova, Kfir Y. Levy, Aurelien Lucchi, Nathanael Perraudin, Ian Goodfellow, Thomas Hofmann, Andreas Krause
Exten-sive experiments show the effectiveness of this measure to rank different GANmodels and capture the typical GAN failure scenarios, including mode collapse andnon-convergentbehaviours.
- North America > Canada > Ontario > Toronto (0.04)
- North America > Canada > British Columbia > Metro Vancouver Regional District > Vancouver (0.04)
- North America > United States > Texas (0.04)
- North America > United States > Massachusetts > Middlesex County > Cambridge (0.04)
- Research Report > Experimental Study (0.93)
- Research Report > New Finding (0.92)
Block-Coordinate Methods and Restarting for Solving Extensive-Form Games
A common approach for solving BSPPs is by using first-order methods, where local gradient information is used to iteratively improve the solution in order to converge to an equilibrium asymptotically.
- North America > United States > Wisconsin > Dane County > Madison (0.14)
- North America > United States > Texas (0.04)