Herke van Hoof

Latest

• Deep Policy Dynamic Programming for Vehicle Routing Problems
• Ancestral Gumbel-Top-k Sampling for Sampling Without Replacement
• Estimating Gradients for Discrete Random Variables by Sampling without Replacement
• Stochastic Beams and Where to Find Them: The Gumbel-Top-k Trick for Sampling Sequences Without Replacement
• Buy 4 REINFORCE Samples, Get a Baseline for Free!
• Attention, Learn to Solve Routing Problems!