Two-Level Softmax Sampling Done Right: Correcting Bias from Size Imbalance and Dispersion
Corrected two-level softmax samplers remove cluster-size and dispersion bias with little overhead.
2 stories, newest first · one card per story
Corrected two-level softmax samplers remove cluster-size and dispersion bias with little overhead.
Theory paper combines differential privacy with fixed-parameter tractable encoders, improving approximation guarantees for combinatorial optimization and proving new lower bounds.
Today 1.5K readers · 2.4K views · 757 reads · all time 189K views