×
Sep 6, 2018 · In this work, we introduce a context-based vocabulary remapping model to reprogram neural networks trained on a specific sequence classification ...
In this work, we develop methods to repurpose text classification neural networks for alternate tasks without modifying the network architecture or parameters.
Feb 25, 2020 · In this work, we introduce a context-based vocabulary remapping model to reprogram neural networks trained on a specific sequence classification ...
Jun 28, 2018 · We demonstrate adversarial reprogramming on six ImageNet classification models, repurposing these models to perform a counting task, as well as classification ...
People also ask
This work introduces a context-based vocabulary remapping model to reprogram neural networks trained on a specific sequence classification task, ...
Adversarial reprogramming is a technique that repurposes a machine learning model, originally trained for a task, to perform a different chosen task.
In this work, we develop techniques to adversar- ially reprogram image classification networks for discrete sequence classification tasks.
This paper demonstrates adversarial reprogramming on six ImageNet classification models, repurposing these models to perform a counting task, as well as ...
We prove that two-layer. ReLU neural networks with random weights can be adversarially reprogrammed to achieve arbitrarily high accuracy on Bernoulli data ...
Adversarial Reprogramming has demonstrated success in utilizing pre-trained neural network classifiers for alternative classification tasks without modification ...