BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Neural Ordinary Differential Equations  - Prof David Duvenaud (Uni
 versity of Toronto)
DTSTART:20180717T140000Z
DTEND:20180717T150000Z
UID:TALK108436@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:We introduce a new family of deep neural network models. Inste
 ad of specifying a discrete sequence of hidden layers\, we parameterize th
 e derivative of the hidden state using a neural network. The output of the
  network is computed using a black-box differential equation solver. These
  continuous-depth models have constant memory cost\, adapt their evaluatio
 n strategy to each input\, and can explicitly trade numerical precision fo
 r speed. We demonstrate these properties in continuous-depth residual netw
 orks and continuous-time latent variable models. We also construct continu
 ous normalizing flows\, a generative model that can train by maximum likel
 ihood\, without partitioning or ordering the data dimensions. For training
 \, we show how to scalably backpropagate through any ODE solver\, without 
 access to its internal operations. This allows end-to-end training of ODEs
  within larger models. 
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR