University of Cambridge > Talks.cam > Machine Learning @ CUED > Backprop through the Void: Optimizing Control Variates for Black-Box Gradient Estimation.

  Browse A-Z

Log in

University Account External (via Google)

🔐 Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹī¸ Information

Backprop through the Void: Optimizing Control Variates for Black-Box Gradient Estimation.

Download to your calendar using vCal

  • UserGeoff Roeder (University of Toronto) Speaker website
  • ClockMonday 27 November 2017, 11:00-12:00
  • HouseCBL Seminar Room.

If you have a question about this talk, please contact .

Gradient-based optimization is the foundation of deep learning and reinforcement learning. Even when the mechanism being optimized is unknown or not differentiable, optimization using high-variance or biased gradient estimates is still often the best strategy. We introduce a general framework for learning low-variance, unbiased gradient estimators for black-box functions of random variables. Our method uses gradients of a neural network trained jointly with model parameters or policies, and is applicable in both discrete and continuous settings. We demonstrate this framework for training discrete latent-variable models. We also give an unbiased, action-conditional extension of the advantage actor-critic reinforcement learning algorithm.

This talk is part of the Machine Learning @ CUED series.

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Seminar Room, Herchel Smith Building, Forvie Site, Robinson Way, Addenbrooke's Hospital, Cambridge CB2 0SR Seminar Room, Herchel Smith Building https://eng-cam.zoom.us/j/82116546594?pwd=RUZQaFI4VndZQ3NRNmtnZmpzTVdtdz09

Other talks

Cambridge AI in Medicine Seminar - July 2026 Max Perutz Lecture - Title TBC Geometry and Generalization: The case of Grokking CONFERENCE: Chemical and material sciences enabled by AI and automation Argentina 2005 Title TBC
 

Š 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity