Recorded Talks

Recurrent Halting Chain for Early Multi-label Classification (20 minutes)

Learning to Selectively Update State Neurons in Recurrent Networks (10 minutes)

Adaptive-Halting Policy Network for Early Classification (3 minutes)