Changing Model Behavior at Test-Time Using Reinforcement Learning

02/24/2017
by   Augustus Odena, et al.
0

Machine learning models are often used at test-time subject to constraints and trade-offs not present at training-time. For example, a computer vision model operating on an embedded device may need to perform real-time inference, or a translation model operating on a cell phone may wish to bound its average compute time in order to be power-efficient. In this work we describe a mixture-of-experts model and show how to change its test-time resource-usage on a per-input basis using reinforcement learning. We test our method on a small MNIST-based example.

READ FULL TEXT

page 5

page 6

research
10/04/2022

Mixup for Test-Time Training

Test-time training provides a new approach solving the problem of domain...
research
09/15/2022

Test-Time Training with Masked Autoencoders

Test-time training adapts to a new test distribution on the fly by optim...
research
04/14/2020

Extrapolation in Gridworld Markov-Decision Processes

Extrapolation in reinforcement learning is the ability to generalize at ...
research
06/10/2021

Scaling Vision with Sparse Mixture of Experts

Sparsely-gated Mixture of Experts networks (MoEs) have demonstrated exce...
research
02/18/2022

Transfer and Marginalize: Explaining Away Label Noise with Privileged Information

Supervised learning datasets often have privileged information, in the f...
research
03/14/2022

Energy-Latency Attacks via Sponge Poisoning

Sponge examples are test-time inputs carefully-optimized to increase ene...
research
02/02/2019

Domain invariant hierarchical embedding for grocery products recognition

Recognizing packaged grocery products based solely on appearance is stil...

Please sign up or login with your details

Forgot password? Click here to reset